The Private Equity Podcast, by Raw Selection

TZP Group adopts AI for deal sourcing and modeling, citing 24x efficiency gains in LBO creation

50 min Listen to the original
Private EquityArtificial IntelligenceOperational EfficiencyDeal SourcingProprietary SoftwareLBO ModelingCost ReductionPortfolio Management

Summary

Jared Berman, Partner at TZP Group, discusses the firm's pivot to AI following a CEO summit hackathon that accelerated adoption among portfolio companies. He highlights the development of a proprietary CRM built in two weeks for $125/month, replacing Deal Cloud to centralize data and reduce SaaS costs. Berman notes that AI reduced LBO model creation time from 72 hours to 3 hours, enabling live scenario analysis. The firm measures success strictly through monetization and cost reduction, using AI to identify add-on targets via public data and standardizing workflows through specific organizational skills.

Episode sections

  1. 0:03 – 2:16

    Introduction and TZP Group Overview

    Jared Berman introduces himself as Partner and Head of AI, Digital Transformation and Growth at TZP Group, a US lower middle market private equity firm with approximately $2.25 billion in AUM and nearly 20 years of history. He outlines TZP's focus areas: property and facility services, wellness, youth enrichment, and branded consumer.

  2. 2:16 – 8:50

    Berman's Career Path to Private Equity

    Berman describes his 15-year background in digital marketing and growth at consumer brands and B2B SaaS companies, including roles at Meta (formerly Facebook) and Gap Inc. He explains that his transition to TZP was driven by relationships and the firm's need for consumer growth expertise beyond just capital.

  3. 8:50 – 13:22

    AI Adoption Catalyst: Claude Excel Plugin

    Berman notes that while AI was a minor line item in TZP's 2026 plan, the launch of Claude's Excel plugin acted as a viral catalyst for adoption among junior investment professionals. This tool significantly reduced the time required to build LBO models, driving a groundswell of interest across the firm.

  4. 13:22 – 18:00

    CEO Summit and AI Hackathon Strategy

    Following news of AI-driven layoffs at Square, TZP pivoted their annual CEO summit to focus on AI. They implemented a 'no-pitch' rule where only CEOs presented use cases and added an AI hackathon. This approach leveraged the competitive nature of PE-backed CEOs, resulting in rapid prototyping of tools like training manuals and sales apps.

  5. 18:00 – 24:55

    Portfolio Company AI Success Stories

    One CEO rebuilt his company's training manual and built a mobile app and field sales tool within weeks of the hackathon. Another CEO used AI for 'vibe design,' increasing the number of SKUs presented to retailers by 100% and attributing over $2 million in projected Q1 2027 revenue to AI-designed products.

  6. 24:55 – 29:08

    TZP Firm-Level AI Skills and Workflow Standardization

    TZP developed 13 'organizational skills' in Claude and ChatGPT to standardize workflows for tasks like NDA review, investment committee decks, and deal screening. They learned that simply asking the AI to perform tasks was insufficient; they had to build specific skills that check for existing work and enforce standardized processes.

  7. 29:08 – 33:25

    Building a Proprietary CRM

    TZP Group replaced Deal Cloud with a proprietary CRM built in two weeks by partner Vlad Guten using AI coding tools. The system costs $125/month to maintain and integrates with Claude, SharePoint, and email to centralize data and reduce SaaS spend.

  8. 33:25 – 36:49

    Measuring AI Success via Productivity and Modeling

    Success is measured by employee productivity gains, such as reducing LBO model creation time from 72 hours to 3 hours. This allows for live scenario analysis rather than static base/downside/upside cases, enabling rapid stress-testing of financial drivers.

  9. 36:49 – 39:27

    AI-Driven Add-On Strategy

    TZP used AI to create an interactive HTML map of franchise locations using public FTD data, identifying top geographic targets for add-ons. This approach built immediate credibility with entrepreneurs by aligning with their existing LOIs and expansion plans.

  10. 39:27 – 43:56

    Monetization and Cost Framework

    TZP adopts Ethan Moloch's framework to measure AI impact strictly through Monetization (M) and Cost reduction (C). AI initiatives must ladder up to existing strategic priorities like revenue generation or margin improvement, rather than creating new priorities.

  11. 43:56 – 50:05

    Continuous Learning and Information Filtering

    Speaker references 'Moonwalking with Einstein' to emphasize maintaining core cognitive skills. He uses AI to synthesize daily newsletters and validates LinkedIn lead-gen prompts by having ChatGPT and Claude assess their validity before sharing with portfolio companies.

Claims worth verifying

TZP Group is a US private equity firm investing in the lower middle market with about $2.25 billion in AUM and nearly 20 years in existence.

Unidentified speaker (speaker_1) ·0:59

How to verify: Company website, SEC Form ADV Part 2A, or press releases regarding AUM milestones. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

Jared Berman joined TZP Group seven years ago in the portfolio growth group.

Unidentified speaker (UNKNOWN) ·1:12

How to verify: LinkedIn profile of Jared Berman or TZP Group press releases announcing his appointment. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

TZP focuses on property and facility services, wellness, youth enrichment, and branded consumer sectors.

Unidentified speaker (speaker_1) ·0:59

How to verify: TZP Group website or SEC Form ADV listing investment focus areas. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

Jared Berman was employee number two at a company that became Inc. 500's number one fastest growing company.

Unidentified speaker (UNKNOWN) ·2:51

How to verify: Inc. 500/5000 historical lists and company press releases identifying early employees. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Jared Berman served as e-commerce marketing manager for the US and later global head of e-commerce marketing on the B2B side at Meta (formerly Facebook).

Unidentified speaker (speaker_1) ·3:11

How to verify: LinkedIn profile of Jared Berman or Meta corporate announcements. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Papa John's and Domino's sell approximately $1 billion of pizza online per year.

Unidentified speaker (speaker_1) ·3:51

How to verify: Company annual reports (10-K) or industry sales data from firms like Statista or Euromonitor. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Primary source

Claude launched an Excel plugin that became a viral sensation among analysts and associates at TZP Group.

Unidentified speaker (speaker_1) ·11:27

How to verify: Anthropic (Claude developer) product announcements or release notes for Excel integration. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Jack Dorsey announced layoffs of 4,000 people at Square, with AI cited as a replacement factor.

Unidentified speaker (UNKNOWN) ·13:41

How to verify: News articles from August/September 2026 reporting on Jack Dorsey's announcement at Square. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

One portfolio CEO wrote 200,000 lines of code and built a mobile app and field sales tracking tool within three weeks of the hackathon.

Unidentified speaker (speaker_1) ·18:52

How to verify: Portfolio company press releases, case studies, or internal TZP reports (if publicly shared). For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

A portfolio company attributed over $2 million in projected Q1 2027 revenue to products designed via AI.

Unidentified speaker (speaker_1) ·24:14

How to verify: Company financial projections or press releases linking AI design to revenue forecasts. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

TZP Group has 13 organizational AI skills, with the most recent being a proprietary CRM built in two weeks.

Unidentified speaker (speaker_1) ·29:08

How to verify: Check TZP Group's website or press releases for announcements regarding their proprietary CRM and AI skills; verify development timeline via internal case studies or LinkedIn posts by Vlad Guten. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

The proprietary CRM costs $125 per month to maintain.

Unidentified speaker (UNKNOWN) ·30:33

How to verify: Review TZP Group's operational cost disclosures in LP reports or internal financials if publicly available; verify against typical low-code/AI development hosting costs. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

AI reduced LBO model creation time from 72 hours to 3 hours.

Unidentified speaker (speaker_1) ·34:14

How to verify: Verify through TZP Group's internal case studies or published benchmarks on AI efficiency in private equity workflows. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

TZP built an interactive HTML map of FTD franchise locations using public data.

Unidentified speaker (UNKNOWN) ·37:15

How to verify: Check FTD's public franchisee directory; verify TZP's map functionality via portfolio company presentations or TZP's website. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Ethan Moloch is a professor at UPenn and an expert on AI usage and measurement.

Unidentified speaker (speaker_1) ·39:27

How to verify: Verify Ethan Moloch's affiliation with the Wharton School at UPenn and his publications on Substack/LinkedIn regarding AI metrics. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

TZP measures AI success through Monetization (M) and Cost (C) frameworks.

Unidentified speaker (speaker_1) ·40:27

How to verify: Check TZP Group's AI strategy documentation or presentations for the M and C framework adoption. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Speaker reads newsletters from Superhuman, AI Fire, and The Neuron.

Unidentified speaker (UNKNOWN) ·45:59

How to verify: Verify subscription status or public mentions of these newsletters by the speaker on LinkedIn or personal website. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify TZP Group's AUM and history via SEC Form ADV filings or their official website.
  • Check the capabilities and limitations of Claude's Excel plugin and similar AI tools for financial modeling.
  • Review public FTD franchise data to assess the accuracy of AI-generated geographic maps for add-on strategies.
  • Compare the total cost of ownership of a proprietary CRM versus enterprise solutions like Deal Cloud, including maintenance and integration costs.
  • Assess the validity of 'vibe coding' outputs by having senior developers or legal teams review AI-generated code and contracts.
  • Investigate the specific 'organizational skills' developed by TZP to understand how they enforce standardized processes in deal screening.
  • Look for case studies or public reports on AI-driven layoffs in tech companies like Square to contextualize TZP's strategic pivot.
  • Validate the claim of $2 million in projected revenue from AI-designed products by reviewing the portfolio company's financial disclosures or press releases.
Full transcriptexpand
0:03
Unidentified speaker (speaker_0): Welcome back to the Raw Selection Private Equity Podcast. Joining us today is Jared Berman, partner, head of AI, digital transformation and growth at TZP Group, a US private equity firm investing in the lower middle market. Today, as you guessed it from the title, this is all about how TZP
0:23
Unidentified speaker (speaker_0): Group are leveraging AI at the firm level and portfolio. This is not a high level discussion. Here is what they are actually doing and how they go about driving adoption of artificial intelligence. Jared, if you can show it as a brief insign to you, please.
0:39
Unidentified speaker (UNKNOWN): Hey
0:39
Unidentified speaker (speaker_1): Alex, thanks so much for having me. I'm I'm incredibly excited to be on the podcast. Loved all your previous guests, and it's a routine weekly listen for me. So thanks again for having me. Uh, By way of background, my name is Jared Berman. I'm a partner at TZP Group. We are a lower middle market private equity firm. We're about to hit our 20th year in existence with about
0:59
Unidentified speaker (speaker_1): two and a quarter billion in AUM. We focus on sort of three core areas, which is property and facility services, wellness, and youth enrichment, and then branded consumer.
1:12
Unidentified speaker (UNKNOWN): I
1:12
Unidentified speaker (speaker_1): joined TCP about seven years ago.
1:15
Unidentified speaker (speaker_1): in our portfolio growth group. It was originally called Portfolio Operating Group, but we have sort of pivoted more focusing on growth. And we can talk a little bit more about that later. And really, my job is to partner with our CEOs, CMOs, chief digital officers on all manners of growth. So that's a combination of how do
1:35
Unidentified speaker (speaker_1): we drive more revenue through marketing? How do we sort of look at greenfield? opportunities for acquisition, new verticals, ways to leverage the software, new clients,
1:46
Unidentified speaker (UNKNOWN): and
1:46
Unidentified speaker (speaker_1): then sort of new tool set. And lately that is really focused on how do we bring AI adoption to the forefront of all of our portfolio companies, which was not a topic actually when we built the 2026 plan for our group, AI was sort of a small line item. in that and has quickly pivoted to be the core
2:05
Unidentified speaker (speaker_1): central tenant of our growth group.
2:08
Unidentified speaker (speaker_0): Marvelous. Well before we dive into the AS stuff, talk to us a bit about your kind of path to private equity because it's a little uh little unusual.
2:16
Unidentified speaker (speaker_1): Sure. So
2:17
Unidentified speaker (UNKNOWN): uh
2:18
Unidentified speaker (speaker_1): it's actually it's it's funny, it's one of the most asked questions I get and probably the number one LinkedIn email I get is hey, how do I get your job? And
2:29
Unidentified speaker (UNKNOWN): the
2:29
Unidentified speaker (speaker_1): truth is, it's it, I'm gonna call it a little bit luck. I'm gonna say a little bit of happenstance, and some of it is you make your own luck along the way. And so my background is I spent the first 15 years of my career in a combination of business development, growth roles in digital marketing, advertising at consumer-led brands and and B2B SaaS
2:49
Unidentified speaker (speaker_1): companies. And through that,
2:51
Unidentified speaker (UNKNOWN): had
2:51
Unidentified speaker (speaker_1): developed sort of a number of great relationships that were born out of very successful growth stories. And so that includes sort of being employee number two at a company that became Inc. 500's number one fastest growing company, taking a legacy retailer and building their e com business from 12 to 56 million in about
3:11
Unidentified speaker (speaker_1): 18 months. And along the way, meeting some wonderful and great people, which ultimately led me to actually take over a vertical within Meta. At the time, it was called Facebook, where I was sort of the e com marketing manager for the US as sort of our head of vertical marketing. And that sort of spanned into a global head of
3:31
Unidentified speaker (speaker_1): e commerce marketing on the B2B side. And that led me to having the opportunity to work with the top 1,000. largest advertisers across Facebook and now Meta. And that was everybody from your pure play, uh, brands like Wayfair and Overstock.com at the time, and brands like
3:51
Unidentified speaker (speaker_1): Wish, but it also was brands that people didn't really think about as sort of traditional e-commerce, which was uh Papa John's and Domino's, which sell a billion dollars of pizza on. online a year. And you don't really think of them because most people would put them in the QSR bucket because their business is selling pizza and restaurants.
4:11
Unidentified speaker (speaker_1): But the method in which they sell and Netflix, again, people think of it as entertainment home video, but everybody in you yourself at some point end up on Netflix.com and need to subscribe. And they are advertising for that. You're going to a landing page, they have the same sort of landing page optimization flows and care about those metrics, lifetime
4:31
Unidentified speaker (speaker_1): value. And even though their product is entertainment, their methodology of acquiring customers is through e-commerce. And so it ended up becoming a much broader set of clients and helping them think through everything from what does great creative look like and how do you tell that story on a vertical format to for
4:51
Unidentified speaker (speaker_1): a while there, everybody was going to a mobile website, and mobile websites were terrible and clunky. And
4:58
Unidentified speaker (UNKNOWN): you
4:58
Unidentified speaker (speaker_1): opened up the keyboard and it went to the letters. And how do you optimize a landing page for very small devices? Obviously, the devices keep getting bigger, and now the iPhone Max is basically like a mini iPad at this point. But how do you keep sort of designing and developing content for that? And so was lucky enough to sort of build those programs for the US and then scale those globally
5:18
Unidentified speaker (speaker_1): across 13 different countries. Um open the first pit Facebook pop up shops. And through that, I had sort of developing my own investing and advising practice for digital disruptors, continued to do that. And for TZP, I got introduced actually through my
5:39
Unidentified speaker (speaker_1): last in-house role on the growth side was a division of Gap Inc., the founder there sold the business to Gap. We actually only overlapped for one week in the role, but he stayed sort of following sort of my growth of the business after he did the transition.
5:56
Unidentified speaker (speaker_1): TZP, as it turned out, had made two consumer investments, was beginning to make more, wanted some help sort of thinking through that.
6:03
Unidentified speaker (speaker_1): And as we'll get into, there's a lot of capital out there. And a key phrase that I'm sure many of your guests have said, and I know they've said it, is sort of what's our right to win? Why does an entrepreneur choose TCP's capital over somebody else's? And often it's the relationships you build at
6:23
Unidentified speaker (speaker_1): the partner level, but uh increasingly as sort of on the operating side. it's sort of what other value add can you bring beyond just the capital, which is a resource that many, many firms have. And so TZP sort of had the foresight which said, hey, we're gonna do more consumer. We need somebody who thinks, eats, and breathes consumer. How do you
6:43
Unidentified speaker (speaker_1): drive consumer growth? What are the key things you need to look at? How do we make better investment decisions? And so seven years ago, I joined leading our sort of digital transformation practice here at TCP. Over the years, that has expanded. And now I spend
6:59
Unidentified speaker (UNKNOWN): about
6:59
Unidentified speaker (speaker_1): a third of my time with our portfolio companies, a third of my time on AI adoption, and a third of our time on sort of how do we find and develop the T C CP P brand brands as well as find great entrepreneurs to invest behind. But the key on how do you get the seat?
7:17
Unidentified speaker (UNKNOWN): Look,
7:17
Unidentified speaker (speaker_1): it's a combination of a couple of things. One, be really good at a functional area. I know you've had many operating partners that are go to market leaders that have been great at that in multiple places. Every port, every P fund is going to have a group of people that are great at go to market.
7:33
Unidentified speaker (speaker_1): Be great at a specific function, number one. Number two is work
7:38
Unidentified speaker (UNKNOWN): for
7:38
Unidentified speaker (speaker_1): a PE backed firm because they never advertise these roles. I mean, you you obviously spend more time than I do, sort of recruiting for roles at PE firms and then sort of their P backed companies themselves. And people want to work with folks that they already know, trust. And everybody else in my group, up until
7:58
Unidentified speaker (speaker_1): myself, had previously worked with the managing partner side by side in a portfolio company. before he started the fund. And so they had already had a twenty plus year track record. I was the first outset hire that had not directly worked for the managing partner, but it was through a relationship, a referral. And I actually spent two years advising and consulting for the fund before I actually
8:18
Unidentified speaker (speaker_1): joined full time. And so it's really is developing that relationship. Um and it's a long game. Like there's a few seats, and so you need to stay around. the hoop long enough to be ready when one of those seats open up to jump in.
8:31
Unidentified speaker (UNKNOWN): And
8:31
Unidentified speaker (speaker_1): then it's you gotta provide value to keep the seat. And that's what I spend the majority of my time on now.
8:39
Unidentified speaker (speaker_0): Appreciate that. Uh that insight. So talk to us about
8:44
Unidentified speaker (speaker_0): the role that artificial intelligence is playing in the in the private equity sphere at the
8:49
Unidentified speaker (UNKNOWN): moment.
8:50
Unidentified speaker (UNKNOWN): Yeah,
8:51
Unidentified speaker (speaker_1): look, it's a it's a great question and and it's something that we think about tremendously. And it's it's a new thought, which I think is something that people have to realize, which is there's obviously a lot of talk about it. It's everywhere. Can't have a go to a dinner conversation and it not come up in some way, shape or form. And I think if I just rewound and I was uh
9:11
Unidentified speaker (speaker_1): looking at August of twenty 2025 twenty five
9:13
Unidentified speaker (UNKNOWN): and the
9:14
Unidentified speaker (speaker_1): question of like what is AI and how do we use it
9:17
Unidentified speaker (UNKNOWN): at
9:17
Unidentified speaker (speaker_1): lower middle market? It was still, I'd say you've seen his fancy Google search. You had access to these at the time
9:25
Unidentified speaker (speaker_1): ChatGBT was really the only game in town.
9:29
Unidentified speaker (UNKNOWN): It
9:29
Unidentified speaker (speaker_1): was still pretty rudimentary. It didn't do things for you. It was
9:33
Unidentified speaker (UNKNOWN): a
9:33
Unidentified speaker (speaker_1): very one-on-one back and forth conversation
9:37
Unidentified speaker (UNKNOWN): that
9:37
Unidentified speaker (speaker_1): got you answers pretty quickly. and helped you think through stuff and
9:42
Unidentified speaker (UNKNOWN): was
9:42
Unidentified speaker (speaker_1): a generally a good thought partner, but it didn't actually do anything for you and it didn't give you extra leverage, I'd say. And that was sort of the way it was all the way through the end of 2025 for most little middle market, which was you were really work focused on
9:59
Unidentified speaker (UNKNOWN): what
10:00
Unidentified speaker (speaker_1): L, is it secure? Is it private? Sort of many of the things that we have to worry about as an SEC registered firm is are we doing everything in a compliant way?
10:09
Unidentified speaker (UNKNOWN): And you
10:09
Unidentified speaker (speaker_1): and I touched on this a little bit, which is you gotta be secure. Like that's sort of step number one. And so a lot of risk averse
10:19
Unidentified speaker (speaker_1): happens in the PE world, right? Like they're never the leading technologists in in finance. And there's a good reason for it. And so for most of 25, it was really, hey, how do we dip our toes in? How do we start to get comfortable in this? And we had an off site and we brought in an AI consulting firm and they helped us sort of, I'd say,
10:39
Unidentified speaker (speaker_1): see a bit of the art of the possible. What if you could run interviews via voice only?
10:45
Unidentified speaker (UNKNOWN): Cool.
10:45
Unidentified speaker (speaker_1): What does that look like? How do you ask questions? How do you submit feedback? And we did a whole little work with 11 laptops. labs and getting comfortable with that. How do you modify a document that you do all the time? How do you build the framework and and ask it to do that? And we played around with that for just for our NDAs, like nothing particularly exciting, but we get thousands of
11:05
Unidentified speaker (speaker_1): NDAs a a year. Spend a lot of time on it. How do you sort of optimize it?
11:10
Unidentified speaker (UNKNOWN): And
11:10
Unidentified speaker (speaker_1): I'd say it sort of was still pretty nascent, even heading in. into 26. And our 26 PGG plan was not AI first. It became AI first. But what started the sort of pivot to AI first was
11:27
Unidentified speaker (speaker_1): Claude comes out and launches the Excel plugin. And that sort of starts to hit, I'll call it the ground and starts to become a viral. sensation amongst analysts, associates, and the folks sort of doing the hardcore LBO modeling where everybody was asking for it. And for a while, most
11:47
Unidentified speaker (speaker_1): of our
11:49
Unidentified speaker (speaker_1): junior investment professionals weren't really leaning into AI. And that sort of starts to come. And that was the lightning rod that I think changed everything, at least in our firm. And I think it was the catalyst that drove.
12:03
Unidentified speaker (speaker_1): hardcore adoption of AI in our portfolio. And from there,
12:08
Unidentified speaker (speaker_1): there was probably three to five of us that were using Chat GBT, had just started getting clawed. And I'd say within two weeks, we started getting requests by every employee for access to Claude. And
12:22
Unidentified speaker (speaker_1): that's just not usually what happens in a PE technology environment. It's it's it's not a groundswell. It's usually like a mandate or a pushdown of like, hey, we're all going to use Teams now. And you get a lot of pushback, and people are like, I hate Teams, or I don't want to use it. Other than email and Excel, like there's really not been a tool that we've
12:42
Unidentified speaker (speaker_1): had a hundred percent adoption or interest in in the seven years I've been here until AI. And once Claude starts working and people started seeing the ability to build within the models and shorten the time it takes to build an LBO, and maybe it's you don't need to get to version 93 like you used to, and you can get away with getting to a perfectly good
13:02
Unidentified speaker (speaker_1): LBO by version 17 now. And the work cutout on that was hours upon hours, if not days, started to sort of hit a groundswell. And the next big inflection point for us was. every year we do a CEO summit where we bring in our CEOs into New York. We host them here for two days.
13:22
Unidentified speaker (UNKNOWN): And
13:22
Unidentified speaker (speaker_1): I'd say it always covered the standard stuff. How do you hire better people, which is why we need folks like Alex and to recruit that because talent is still the number one performance indicator of are we going to have a great outcome? Is who sits in those seats. But there's only so much you could talk about. putting in great people in the seats.
13:41
Unidentified speaker (UNKNOWN): How do
13:41
Unidentified speaker (speaker_1): you improve EVITA? How do you work on margin and margin expansion and revenue generation? So those are always the core topics. We're planning the agenda. It's probably three and a half weeks out from the summit. We have a working agenda. Everyone's flights are booked. And Jack Dorsey puts out his missive about laying off 4,000 people at Square.
14:01
Unidentified speaker (speaker_1): And the that they're at least from the article, gonna be replaced by AI. And
14:08
Unidentified speaker (UNKNOWN): I
14:09
Unidentified speaker (speaker_1): sat there with another one of my the operating partners, Michelle named Came Estero, incredibly accomplished. And I said, We got to throw, we got we looked at each
14:15
Unidentified speaker (UNKNOWN): other
14:16
Unidentified speaker (speaker_1): and we're like, we got to rip up this entire agenda. Like, this is all anybody's gonna be talking about is AI, AI, AI. And in the span of three weeks, we ripped up our agenda, we've rebuilt We it. We sourced new speakers and we built an entirely new program. And
14:32
Unidentified speaker (speaker_1): we didn't know what we were gonna, how it was gonna go. We had no idea. And what did we get right? We got a couple of things really right that I think fur catalyzed our portfolio companies. Number one,
14:45
Unidentified speaker (UNKNOWN): we
14:45
Unidentified speaker (speaker_1): had a no no-pitch pitch rule. So while we brought in experts, we weren't there to have them pitch their AI. tool of how they're going to help you do anything. It was really, and the reason we got the no pitch was we only let other CEOs present. So regardless of what your tool was, you had to be the CEO of that company. And the key
15:05
Unidentified speaker (speaker_1): questions we asked them is,
15:07
Unidentified speaker (UNKNOWN): how
15:07
Unidentified speaker (speaker_1): are you as a CEO that has all the same challenges as people in this room? You have employees, you have boards, you have investors. you have financial needs that you have to hit, you have revenue projections, you have all the same challenges. You are probably just further along since you're developing an AI company in the usage of AI. And so we really said,
15:28
Unidentified speaker (speaker_1): how are you a better CEO because you have act you're further along in AI? An And what are the tools you're using? What are you reading? How are you learning? How much time is it taking? to really demystify AI to everybody? Which is you can learn it very quickly if you spend time on it.
15:47
Unidentified speaker (UNKNOWN): Yep.
15:48
Unidentified speaker (speaker_1): So that was sort of our attendee list, was like you had our speaker list, was you had to be a CEO.
15:53
Unidentified speaker (UNKNOWN): The
15:53
Unidentified speaker (speaker_1): second thing we did, and this was sort of something we took from Meadow, which was they did a hackathon every six months. How do you build new products? How do you think about it? And as we ourselves were now two months into the TZP journey of really getting better and better ourselves, we recognized that if you don't actually use the tool,
16:11
Unidentified speaker (UNKNOWN): you
16:11
Unidentified speaker (speaker_1): don't get better. And so we added the day with an AI hackathon, which was we spent
16:19
Unidentified speaker (UNKNOWN): a
16:19
Unidentified speaker (speaker_1): better part of six or seven hours inspiring you, teaching you, showing you the connectors, explaining how they work. very little things that you can do that can add real value to your day-to-day life as a CEO. You now have 45 minutes to do something.
16:35
Unidentified speaker (UNKNOWN): And
16:36
Unidentified speaker (speaker_1): at the end of it, you got to present it to all the other CEOs.
16:40
Unidentified speaker (speaker_1): And
16:42
Unidentified speaker (speaker_1): you've been enough around enough PE back CEOs that there's an a innate competitive streak in anybody who's a CEO at a PEC. PE backed company, which is I don't want to be the guy that or girl that sort of puts something out there and it's
16:57
Unidentified speaker (speaker_1): not so great. And so I think it got their competitive juices flowing a little bit. And it was one of those things where like when you were back in grade school and the teacher's like, open your notebooks and the test starts now and everybody just goes quiet and you just sort of hear the scribbling. We said you got 45 minutes, we'll walk around and help. and advise. And if you need connector issues or
17:17
Unidentified speaker (speaker_1): anything like that, we'll walk you through it. But like you got 45 minutes to come up with something. And you then went to like pin drop silence. Like people were like, I gotta do something and I gotta show it off and I gotta be proud of it.
17:30
Unidentified speaker (UNKNOWN): And
17:31
Unidentified speaker (speaker_1): we weren't looking for somebody to paint the Mona Lisa here. We're like, come up with something that's really valuable. And we had some amazing things done in 45 minutes. One of our CEOs that morning subscribed to Claude, which was the only requirement, was you had a paid Claude account to show up. That morning they did it. Like while we were eating breakfast, he hadn't done it, he
17:51
Unidentified speaker (speaker_1): did it.
17:52
Unidentified speaker (UNKNOWN): He
17:52
Unidentified speaker (speaker_1): then was like, Look, I I don't feel like uh the knowledge as a founder has sort of always stayed in my head, has really sort of been unlocked, even with our head of sales or our technology. It just sort of is still really local to me. And this is after 15 plus years of running the company. He's like, I really think I gotta rebuild the operating
18:12
Unidentified speaker (speaker_1): training manual for our staff. And he literally just started typing all the things that he felt like people at the company just weren't getting out of his head into it and said, rebuild the training manual and gave uploaded the existing one. typed it out, and within 45 minutes, he had rebuilt the
18:32
Unidentified speaker (speaker_1): intro to the company and how to tell our story in 45 minutes based off of everything that was in his head. And I think for him it was a light shed moment. We then sort of learned that that CEO within three weeks had written three, had become so into it. it's It was like we plugged him them into the matrix, had written 200. 200,000 lines
18:52
Unidentified speaker (speaker_1): of code and replaced the CTO, built a mobile app, and built a field sales tracking tool for the field employees, all after that.
19:03
Unidentified speaker (UNKNOWN): Now
19:04
Unidentified speaker (speaker_1): the ROI on that, which I know we're going to talk about, is look, there's obviously some people savings in that area,
19:10
Unidentified speaker (UNKNOWN): but
19:11
Unidentified speaker (speaker_1): there's also sort of efficiency gains, there's training and increased sales conversion rate because you're now sort of all selling the same way. You're getting the highest and best use of every sales employee
19:23
Unidentified speaker (UNKNOWN): on
19:23
Unidentified speaker (speaker_1): any CRM. I know you sort of mentioned you have a CRM for your own
19:27
Unidentified speaker (UNKNOWN): for
19:27
Unidentified speaker (speaker_1): your own business. The hardest part of the CRM is people putting stuff into it and using it. That's the hardest part. So the fact that he was able to develop an app, because of what the field was going through, he knew better than anyone. anybody else who's going to develop it. What are they going to do? What aren't they going to do? And what fields do they really not care about? And like I'm just going to
19:45
Unidentified speaker (UNKNOWN): get
19:45
Unidentified speaker (speaker_1): bad data on. And so he laid that all out and built it and rolled it out. And it's been incredibly successful. We've had other,
19:53
Unidentified speaker (UNKNOWN): we
19:53
Unidentified speaker (speaker_1): had other CEOs sort of optimize their shipping routes because they had to move a lot of boxes around. How do you sort of do that more efficiently?
20:02
Unidentified speaker (UNKNOWN): And then
20:02
Unidentified speaker (speaker_1): we just had a lot of personal stuff, which is how do I set up my day for better success? How do I manage my calendar better? And those things are a little bit harder to measure. But in a busy CEO world, like
20:14
Unidentified speaker (speaker_1): being able to bucket your time so you're not having those switching costs every couple of minutes of,
20:21
Unidentified speaker (UNKNOWN): oh,
20:21
Unidentified speaker (speaker_1): I had a call. Then I was sort of trying to do some real strategic thinking. And now so we really was sort of setting up their own calendar of like, here are the things that I need to do every week, figure out the best way to optimize my one on ones, my strategic thinking, my reporting out to the board, my store visits or whatever it might be. And so I think a lot
20:41
Unidentified speaker (speaker_1): of people spend a lot of time doing that. And then we shared it all
20:44
Unidentified speaker (UNKNOWN): out. And
20:45
Unidentified speaker (speaker_1): what was so great about it is we really improved people's confidence. When we sort of started a day. the day, we asked people on a scale of one to five, how how confident are you on AI? It was sort of five being I'm Sam Altman and zero being what's AI. And we were in the 1.9 to 2 and a half range, I would say, was sort of the the
21:05
Unidentified speaker (speaker_1): average. When people left, we were over four.
21:09
Unidentified speaker (UNKNOWN): Good.
21:09
Unidentified speaker (speaker_1): So it was a massive increase of confidence in of themselves. And that allowed them to bring it back to the companies. And I think the second big inflection point that we've had with AI was sort of number one was you got to get people excited and using
21:25
Unidentified speaker (UNKNOWN): it.
21:25
Unidentified speaker (speaker_1): The second big thing is after that happens, a lot of stuff just gets created. Like a people are just doing stuff, right? You're like, I got AI and I'm just gonna do things with it. I'm gonna ask it a lot of questions. I'm gonna load data into it, I'm gonna and get a lot of outputs. And I'd say this that is an incredibly important time period.
21:45
Unidentified speaker (UNKNOWN): We're
21:46
Unidentified speaker (speaker_1): not gonna call that sort of three to six months of the a lot of stuff gets produced. Much of it is not high value in terms of driving nav, even to improvements, but it is really the core foundational of hey.
22:02
Unidentified speaker (speaker_1): if they don't spend time learning, making mistakes, trialing it, understanding what it can do, you actually don't get to phase three. And phase three is the most exciting part, which is people are now, I'm gonna call it nearly fluent in what they can do. They might not be using Kodaks or odd code, but they're using cowork.
22:22
Unidentified speaker (speaker_1): They know that it can automate some stuff for them, they know that it can read, write. scrape, any sort of those skills. Now the part is okay, how do we tie that to core nav? And that
22:34
Unidentified speaker (UNKNOWN): was
22:35
Unidentified speaker (speaker_1): sort of how we frame sort of our own planet TCP, which was the first six months was really drive adoption, adoption, adoption, adoption internally and at the porcos. The back half of this For year is how do we get five to seven key nav wins based off of the fact that people have now had this adoption? And that's what we're we're
22:55
Unidentified speaker (speaker_1): excited about for the back half of this year. And we're seeing it in massive ways. I talked about sort of our first CEO and what he's doing for his field employees
23:04
Unidentified speaker (UNKNOWN): and
23:04
Unidentified speaker (speaker_1): sort of the employee optimization. We have another CEO that I refer to her as is sort of the the first of vibe designing, which is was how do you shorten the product design lifecycle?
23:15
Unidentified speaker (speaker_1): Go from
23:17
Unidentified speaker (UNKNOWN): it
23:17
Unidentified speaker (speaker_1): takes me six months to draw, innovate, send it to a factory, then I have to present it to a retailer, then I get some feedback, then I go and distribute it. Now
23:28
Unidentified speaker (UNKNOWN): one
23:29
Unidentified speaker (speaker_1): of our CEOs attributes 80% of the design work to being done in AI and it allows them to bring to retail meetings a hundred percent more available. SKUs to the retail meeting. The business has exploded in the last six months since they've done that. Because they used to walk in because there was a people constraint. Somebody can
23:49
Unidentified speaker (speaker_1): only Adobe Photoshop X amount of baby products
23:54
Unidentified speaker (UNKNOWN): and
23:55
Unidentified speaker (speaker_1): then sort of get it to a factory and be like, can we actually make this? Now they're like, hey, we know what the constraints are because we've already built so many. SKUs. So as long as we build within those constraints, we're pretty sure the factory can make it. Now it's a matter of are we getting the trends right? And you get the trends right when you give a lot more people a lot more options to pick out what the
24:14
Unidentified speaker (speaker_1): trends are going to be. And so they used to walk in with three, five, seven different patterns. Now they walk in with 25 At different patterns and are like, which one do you want? And because of that, they are attributing over two 2 million of 2022. 2027 Q1 revenue to products that were designed via AI.
24:32
Unidentified speaker (speaker_0): Wow. What a what a journey. What a journey. So
24:36
Unidentified speaker (UNKNOWN): a
24:36
Unidentified speaker (speaker_0): lot happening there, a lot of use cases, lots of different things happening at private equity level and portfolio level. We spoke you spoke a lot about the portfolio and then how you've begun to do that at private equity level. You mentioned right at the start the NDA. How are TZ? TZP at the P firm level also leveraging artificial intelligence?
24:55
Unidentified speaker (speaker_1): It's a great question. So
24:57
Unidentified speaker (UNKNOWN): the
24:58
Unidentified speaker (speaker_1): way we started doing it was we we looked at what are the things that people are spending a lot of time on that aren't particularly the most value added, but are a key part of the P workflow. Every
25:09
Unidentified speaker (UNKNOWN): P
25:10
Unidentified speaker (speaker_1): person who's listening to this has the same sort of things that are happening, right? Which is you get an email from a banker. first thing they do is they send you the teaser. The teaser excites you enough, they're like, you respond back, interested. Next thing that comes in the email is the NDA.
25:24
Unidentified speaker (UNKNOWN): At
25:25
Unidentified speaker (speaker_1): this point, every firm has edited um I mean, thousands upon thousands of NDAs. There's things that we care about, there are things that we don't care about,
25:34
Unidentified speaker (UNKNOWN): but
25:34
Unidentified speaker (speaker_1): all these tools can be edited. So what we have begun developing at TCP is what we call our organizational skills, which is something that's available in both Claude and Chat GPT. And really, it is sort of these systems that say, hey, if you have to do one of these core functions, instead of you reinventing the wheel every time, we've
25:54
Unidentified speaker (speaker_1): built a skill for you. It's deployed in your instance and you should use it. So we have that across. We have We've built 13. built 13 organizational skills. So there are ones for NDA, there's ones for a heads up memo, there's ones for the foundations and skeletons of an IC deck. We then have it for
26:14
Unidentified speaker (speaker_1): obviously NDA, which we talked about. We have a SIM analyzer, which sort of is another sort of key area, which is does it sort of fit our investment thesis? Does it hit sort of our and then gives it a rating? and everything we do still still has human eyes, but it sort of is a first stab at a recommendation of like pass fail. And if
26:34
Unidentified speaker (speaker_1): it's a pass, what level of pass is it?
26:37
Unidentified speaker (UNKNOWN): So
26:37
Unidentified speaker (speaker_1): we built that out. And the truth is we thought it was so smart and working. And then 90 days later,
26:46
Unidentified speaker (UNKNOWN): we
26:46
Unidentified speaker (speaker_1): were like, hey, is this actually happening? Because we would go into meetings and somebody would have a format that looked one way, somebody have a format that looked another way. And we're like,
26:53
Unidentified speaker (UNKNOWN): we
26:54
Unidentified speaker (speaker_1): like built this and we like pushed it, and it's supposed to work because it's an organizational scale. And here's what we found out. And this was actually really interesting, which is the things have become so smart, so smart, that if you just say, Hey, I need to edit MDA, guess what? Claude can edit MDA NDA all day long without any problem.
27:13
Unidentified speaker (UNKNOWN): How do
27:14
Unidentified speaker (speaker_1): you analyze? this PDF? Guess what? Blood will analyze any PDF you give it. What it wasn't doing, it was doing it the TZP way.
27:21
Unidentified speaker (UNKNOWN): And
27:22
Unidentified speaker (speaker_1): this is what we learned the hard way, which was just asking it to do something doesn't actually invoke a scale.
27:28
Unidentified speaker (UNKNOWN): So
27:28
Unidentified speaker (speaker_1): we then had to go back and actually rewrite the rules for how does Edderon's Claude actually work and Chat GBT. And so what we did was we actually built,
27:37
Unidentified speaker (UNKNOWN): of course,
27:38
Unidentified speaker (speaker_1): another skill. which is before any of the AI systems will do any work, it now does three things. Number one, it checks if there's a skill that we have designed and developed that covers what you're asking for to
27:51
Unidentified speaker (UNKNOWN): do.
27:52
Unidentified speaker (speaker_1): That's number one. Number two, has somebody actually already done this? Right? Because on any one of those emails, seven, four to seven people are copied in at our firm.
28:03
Unidentified speaker (UNKNOWN): A
28:03
Unidentified speaker (speaker_1): little bit of a jump ball, who's going to do it first. Hopefully it sort of gets tracked.
28:09
Unidentified speaker (UNKNOWN): But
28:09
Unidentified speaker (speaker_1): it wasn't always, and people were editing the same NDA. So now it says,
28:13
Unidentified speaker (UNKNOWN): Do
28:14
Unidentified speaker (speaker_1): we have a skill? Yes or no? If it's if it's a yes, it asks you, Do you want to use this skill? So you have to affirmatively apply the yes, I want to do it. Number two is it checks,
28:24
Unidentified speaker (UNKNOWN): did
28:25
Unidentified speaker (speaker_1): somebody already do this? And it reads. our SharePoint drive to check if this has already been done.
28:30
Unidentified speaker (UNKNOWN): And
28:30
Unidentified speaker (speaker_1): so it'll find the company name, it'll find the project name, whatever it is. It'll go into our SharePoint folder section. It knows where to look. We've mapped it. And then it says, this hasn't been done. Do you want to do it now? So step one, what skill do you want to use? Step two hasn't been done. Step three, which is really great, says, is this something you're going to do repeatedly?
28:51
Unidentified speaker (speaker_1): And after I finish doing this, should I actually turn this into a skill if it doesn't exist?
28:56
Unidentified speaker (UNKNOWN): Yeah.
28:57
Unidentified speaker (speaker_1): And so it starts to create this sort of compounding effect because what we were getting was a lot of work, but we weren't actually getting the compound compounding impact internally. So we have 13 skills now.
29:07
Unidentified speaker (UNKNOWN): And
29:08
Unidentified speaker (speaker_1): the most recent one we did, which was really exciting, was we built TZP CRM, which is our own proprietary. in-house CRM tool. And that was a move away from Deal Cloud. And
29:21
Unidentified speaker (speaker_1): many of I'm sure your your listeners or clients have used Deal Cloud. And there's a few of them that are P specific. And Deal Cloud's wonderful for very from for many years. We've were happy clients,
29:35
Unidentified speaker (UNKNOWN): but
29:36
Unidentified speaker (speaker_1): we sort of got to the point where like, are we getting enough value? And I think this sort of comes to many of the things you've talked about in the SAS pocalypse and the things that are happening in the broader ecosystem, which is
29:47
Unidentified speaker (speaker_1): are there things that we want that we don't have in our current CRM that would make us more efficient? And and is it also a great sort of proof point to our CEOs itself that like, hey, if we can build something, you could build something, right?
30:03
Unidentified speaker (speaker_1): And we decided that we were going to build our own CRM. And one of the other partners in the firm loves great tinker, um, guy named Vlad Guten. He was like, I'm gonna do it. And
30:17
Unidentified speaker (UNKNOWN): we
30:18
Unidentified speaker (speaker_1): scoped out what did we like in Deal Cloud, what do we like in every C other CRM. We put together our own sort of mini internal RFP of what we'd want our systems. system to do.
30:28
Unidentified speaker (UNKNOWN): And
30:28
Unidentified speaker (speaker_1): then he we vibe coded it to a full CRM in two weeks.
30:33
Unidentified speaker (UNKNOWN): And
30:33
Unidentified speaker (speaker_1): we rolled it out to all the employees. It has a web interface. It will soon have an Outlook plugin to do the one click ad that we're vibe coding. But more importantly, it allows us to sort of clean the data, own the data. It gives everybody their own centralized dashboard, which a little bit. And the
30:53
Unidentified speaker (speaker_1): sum total per month is $125 a month to maintain it. Not including, I'm going to call it the human hours to sort of update the bugs and any features, but from a pure dollars,
31:05
Unidentified speaker (UNKNOWN): which
31:06
Unidentified speaker (speaker_1): obviously is an opportunity cost when people are doing one thing versus another, but I'm going to gonna exclude that for the moment. But from a pure how many dollars were going out versus how many are going out now.
31:15
Unidentified speaker (speaker_1): Look, we get those those are things that we have to deal with with our LPs, our funds, to be able to go to them and say, hey, we're treating your dollars incredibly preciously, and our costs to run the administration of the firm are going down because we're building proprietary tools that are gonna going to make us better is
31:35
Unidentified speaker (speaker_1): really important.
31:37
Unidentified speaker (UNKNOWN): And
31:37
Unidentified speaker (speaker_1): so we built it, we launched it. and it's been incredible. And we continue to add more and more features to it. And we're enriching the data now. And because you are not spending money on one thing, you can do other things, right? We didn't get the ability to add an Apollo or a Rocket Reach or a Clay and sort of really
31:56
Unidentified speaker (speaker_1): enrich the data because you and I talked about this previously, but like people change roles and it happens more often. And some of the systems are not particularly great at updating that. And this is in a way that we traditionally haven't really done it that way, but now we can. I mean, it gives you a daily what do you need to do, and it's live, it's editable at any one
32:16
Unidentified speaker (speaker_1): point. And the last piece we did was we build the connector straight into Claude. So now anytime you want, you can literally just type please. log this email in CRM. Andor please find me all the 10 most recent deals of bagel shops
32:36
Unidentified speaker (speaker_1): we looked at as like an example. And it will go find all that. It will put into Claude. You can then ask for where all these files hosted. It syncs with our SharePoint.
32:45
Unidentified speaker (UNKNOWN): And
32:45
Unidentified speaker (speaker_1): so now it is all connected into one ecosystem, which is Claude cloud, SharePoint email.
32:52
Unidentified speaker (speaker_1): All integrated. And it wasn't like that before. And because it was purposely built for
32:56
Unidentified speaker (UNKNOWN): us,
32:57
Unidentified speaker (speaker_1): it really sort of has these added features. And that's probably one of the most used skills now is the CRM skills that we built into cloud.
33:06
Unidentified speaker (speaker_0): Wow. What a project. So looking at that and with AI, I think a lot of people have gone down the rabbit hole and found that there's not as much value as they'd thought. It ends up costing more, you know, the old token chat and other bits and pieces. So at both firm level and portfolio level, how are you measuring AI success?
33:25
Unidentified speaker (UNKNOWN): So
33:25
Unidentified speaker (speaker_1): I think there's there's two things. One is I think people have felt they had to
33:31
Unidentified speaker (UNKNOWN): sign
33:31
Unidentified speaker (speaker_1): up for bigger and bigger usage is overall.
33:35
Unidentified speaker (UNKNOWN): And
33:35
Unidentified speaker (speaker_1): that whether that's enterprise, business, we personally use the business. the small business accounts for everybody. It keeps our token fees down. And then we have added coverage if people need more tokens. And so we have no problem overspending, but we found that sort of the cost benefit analysis on some of the larger tiered business
33:55
Unidentified speaker (speaker_1): plans were well above what we needed need it right now. And so I think a lot of folks I would say jump right into thinking they need enterprise I would always tell them, hey, before you go to enterprise, sort of see whether or not the small business is good enough for you. If it's not, of course, good do what you need to do.
34:11
Unidentified speaker (UNKNOWN): So
34:11
Unidentified speaker (speaker_1): we think about it in two ways. Number one is
34:14
Unidentified speaker (speaker_1): employee productivity at the firm
34:18
Unidentified speaker (UNKNOWN): is
34:18
Unidentified speaker (speaker_1): number one, which is, hey, if it used to take you three have days to make an LBO, now there's an LBO skill and it builds you a fully reactive HTML. formatted LBO.
34:29
Unidentified speaker (speaker_1): And if it used to take you, let's just call it 72 hours of real work, and then now it takes you, let's just say, three to get you like a grade V1, the work isn't done there. And I think that's the next part of the conversation that people need to have, which is just because you've got the output, that was the 72 seven to hours of work, now it's spend
34:49
Unidentified speaker (speaker_1): real time, spend another. three hours on understanding all the drivers, slide the scale on the
34:57
Unidentified speaker (UNKNOWN): the
34:58
Unidentified speaker (speaker_1): the debt rate, slide the scale on customer acquisition costs, slide the scale on employees, and really understand like if I move one metric, because it's now so easy to move one metric,
35:10
Unidentified speaker (UNKNOWN): what
35:10
Unidentified speaker (speaker_1): does that do for you?
35:12
Unidentified speaker (UNKNOWN): Um
35:13
Unidentified speaker (speaker_1): and really understand that because it it used to be sort of just like get it done. And then you ended up with sort of like three cases. You ended up with sort of like the base case, the downside case, and the upside case. Now you have we have live cases, right? Which is like, what if three things happen simultaneously? Which is we get a new client,
35:33
Unidentified speaker (speaker_1): we have to spend the cash to produce goods, and
35:37
Unidentified speaker (UNKNOWN): we
35:37
Unidentified speaker (speaker_1): get paid. 20 days later than we expected. Well, that's a pretty niche sort of a
35:45
Unidentified speaker (speaker_1): case that you normally wouldn't run. But now we've seen it. We've had our portfolio companies win a new client. Okay, great. Super exciting. A lot of revenue. A lot of costs right up front to generate that revenue. What happens? Now we can actually very quickly run that live. Like you open up the HTML, you slide it over, and immediately shows
36:05
Unidentified speaker (speaker_1): you. And so you can run all that. And so it's really helpful to understand: hey, if this happens, here are what we need to be prepared for.
36:13
Unidentified speaker (UNKNOWN): These
36:13
Unidentified speaker (speaker_1): are the metrics that are going to move. And you couldn't do that before. So I really view it as like you save a lot of time, but you got to take some of that time that was saved and really immerse yourself in what the output is. You're not done when the output comes out. Then it's the what can you learn and how do you become more disciplined in helping work with our port codes and our CEOs on
36:33
Unidentified speaker (speaker_1): if these things are happening, these are the downstream effects. And we know it because we have the ability to forecast and play with that in the model, where we normally wouldn't have. We had to build a new case or a new et cetera. And now that's live. And so I think that's sort of number one is time safe. saved. Then there's the hey,
36:49
Unidentified speaker (UNKNOWN): what
36:49
Unidentified speaker (speaker_1): can we do that we couldn't do? previously? So we are looking at a four-wall business. It's a franchise space. Uh a key piece of it, like many sort of P things is, oh, well, let's let's add on. Let's do add-ons. Let's add 20 more locations in our geographic area because we already have the the footprint, we have the core
37:09
Unidentified speaker (speaker_1): operating team. Then we can have sort of the retail staff that we can move around to locations. Great.
37:15
Unidentified speaker (UNKNOWN): Okay,
37:15
Unidentified speaker (speaker_1): well, who owns what? Where are they from? The nice thing is this is public. So we went to the franchise, the FTD, the franchise that they put out. It lists all the franchisees, where they are, et cetera. And we built an interactive HTML map of every location, In roughly size, that we can sort
37:35
Unidentified speaker (speaker_1): of figure out who owned them, where were they? And it actually built a geographic target list for us right away. And when we showed that to the entrepreneur and he was like, Yeah, I want to do add-ons, which one are the right ones? We unbeknownst to us, we put up our map and
37:53
Unidentified speaker (UNKNOWN): we
37:53
Unidentified speaker (speaker_1): said, Look, these would be the top five that we would go after right away. Like within the first week of closing, we should, and even while we're under LOI, we should start contacting these. And he goes,
38:04
Unidentified speaker (UNKNOWN): and
38:04
Unidentified speaker (speaker_1): we got immediate credibility because unbeknownst to me, he was operating as though, hey, if I don't get a deal done, I still got to do this on my own. He was under LOI with four of our top five that we identified.
38:16
Unidentified speaker (UNKNOWN): And
38:16
Unidentified speaker (speaker_1): so it's instant credibility that we built this map. We said, based off of where your footprint is, we think these are the most four geographically desirable locations to go to. And he goes, I'm actually in,
38:28
Unidentified speaker (UNKNOWN): I
38:28
Unidentified speaker (speaker_1): actually have this under LOI right
38:30
Unidentified speaker (UNKNOWN): now.
38:30
Unidentified speaker (speaker_1): So it immediately showed that like, hey, we did the work. We understood his business. We understood the radius of where his employees were, and we had the ideas that he already had, which should have said, okay, they speak the same language as me. They understand it. Now do I think over a year ago you could have taken the FTD list and put it in Excel
38:51
Unidentified speaker (speaker_1): and sort of like done a Google Maps thing, which is like not nearly as interactive as what we did, which is like it clicks on it, opens up who owns the FTD, it then has a link out to the LLC registration page to see if we could find contact information. You couldn't build all of that at one time. So now you have this ability to be much more efficient in add-ons and
39:11
Unidentified speaker (speaker_1): different areas that I I think you couldn't do. And it's hard. to sort of say, like, what's the value of something you couldn't do
39:17
Unidentified speaker (UNKNOWN): on
39:17
Unidentified speaker (speaker_1): the portfolio company side, which is sort of the second part of your question. We view it as sort of there's a great speaker. He's a professor at UPenn,
39:27
Unidentified speaker (speaker_1): Ethan Moloch. Maybe I think you probably have heard of him, you follow him, but he's writes on Twitter, he writes on Substack, Stack writes right on LinkedIn, and he's is one of the foremost experts on sort of uh AI usage and measurement. And he has sort of really built it. And I think learning from other people and being Spark, Vista Equity has sort of done a really good job at given they are
39:47
Unidentified speaker (speaker_1): very focused on one vertical, they've sort of really figured out sort of this AI mapping tool and framework. And so we have been adopting it ourselves, which is what gets measured is what gets done, right? And what we sort of want to improve. upon is things that are already important. So the goal here isn't just because you have
40:07
Unidentified speaker (speaker_1): AI to go do a lot of stuff. It's how does AI make you better at the things you've already identified as a key strategic initiative in that hundred-day plan, in that mid-hold review, and really sort of look at that and say, how does AI roll up? And sometimes it doesn't. And sometimes it does. And so we've built yielded it as two things, M and C.
40:27
Unidentified speaker (speaker_1): One, monetization. How does AI help you on any of your core initiatives drive better monetization?
40:35
Unidentified speaker (UNKNOWN): And
40:35
Unidentified speaker (speaker_1): the C stands for costs. How does AI help you reduce costs? And is that could be software, that could be human, that can be
40:45
Unidentified speaker (speaker_1): any sort of expenses that the company has. How do you reduce those? And so one is revenue generation, one, I'll call it a sort of margin. increasing, right? As you spend less, your margin goes up. Those are ultimately the real building blocks of fundamentals of any business, right? Which is you need to drive revenue and you want it to be the most profitable revenue you can drive. So we're really focused now on
41:05
Unidentified speaker (speaker_1): everything we do for AI is to align with the portfolio company executive team, with the deal team, and say, what are your core areas of focus? Can I help you or not?
41:17
Unidentified speaker (UNKNOWN): And
41:17
Unidentified speaker (speaker_1): it's okay if it's a not. Like this is just another tool. It's not going to do everything for you. It's not a magic bullet.
41:23
Unidentified speaker (UNKNOWN): But
41:24
Unidentified speaker (speaker_1): if it can, here are the things to do. Here's how we can help. Here are the resources. Here's how we can help your team build it internally. Here's how we can help you find the right outside third third-party party consultants. Here's how TDP's internal team could be supportive of that as well and help. project management. And it's really comes down to the M and the C. What drives monetization?
41:44
Unidentified speaker (speaker_1): What drives cost? And everything needs to ladder to that. And I think that's the part that I'm calling it, which we talked about if I go back a little bit, was sort of phase three, which was you now know that it can do a lot of things.
41:57
Unidentified speaker (UNKNOWN): Now
41:57
Unidentified speaker (speaker_1): you need to harness it for the most important things.
42:01
Unidentified speaker (UNKNOWN): And
42:01
Unidentified speaker (speaker_1): it's sort of like the eye. iPhone and email when everyone sort of started getting their work emails on their phone. Okay, you were more accessible. Did it make you do better work because you were more accessible? Not always. It was just sort of you can reach people much easier and they're always on. It doesn't mean they were always doing great things because they had email
42:21
Unidentified speaker (speaker_1): in their pocket.
42:22
Unidentified speaker (UNKNOWN): The
42:22
Unidentified speaker (speaker_1): fact that you have AI doesn't mean that you were doing the most valuable things with AI.
42:27
Unidentified speaker (UNKNOWN): But
42:27
Unidentified speaker (speaker_1): I think it's okay. I think it was because there that phase two of really learning, understanding, and playing with it will make the phase three much more successful because you're not wondering if it can do it. Now it's really how do we do it the best way possible to get the maximum in terms of monetization or the maximum in cost
42:47
Unidentified speaker (speaker_1): benefits? And that's really sort of, I'd say, the back half. for us. And everything's being measured of hey, if it's speed to lead, what was our speed to lead before we use something related to AI? What is it after? What do we spend on outsourced technology tools? What are we spending now? Do we need three people
43:07
Unidentified speaker (speaker_1): to do something, or can you And deal with one plus some automations and really sort of marrying everything up so that everything ladders up to an M1, M2, M3, C one, C2, C three. But it's really got to be around what are the highest and most important priorities existing for the company, not create new priorities. It's got to
43:27
Unidentified speaker (speaker_1): ladder up to what do we already know is the key drivers for the business. And then how does AI leverage up?
43:32
Unidentified speaker (speaker_0): Following up on that kind of constraint, what's the uh biggest constraint if result bring the largest return on investment? Trying to fix that, not just trying to leverage uh certainly leverage AI. Well, what a what a podcast. What an what a journey from how you guys are adopting to all the way through. What's what's a book, framework, or resource that's had the biggest influence on your career,
43:52
Unidentified speaker (speaker_0): aside from uh the coin term of artificial intelligence?
43:55
Unidentified speaker (speaker_1): Sure.
43:56
Unidentified speaker (UNKNOWN): I
43:56
Unidentified speaker (speaker_1): mean, one of my favorite books, which I think is an under-the-radar one, is I love a book called Moonwalking with Einstein. Have you heard of this?
44:03
Unidentified speaker (speaker_0): No, I have
44:03
Unidentified speaker (UNKNOWN): not.
44:04
Unidentified speaker (speaker_1): So it was about a reporter that entered the world memory championship. And the question was, could you train yourself to be a better at mem remembering things? And which is like an odd thing in juxtaposed AI, which sort of, I think just yesterday they talked about the new memory and how flawed members more of your
44:24
Unidentified speaker (speaker_1): sessions. Um, but I do think.
44:27
Unidentified speaker (speaker_1): as people use more and more technology, I have kids, they're using it more. You do sort of worry about like our core things that you used to have to do. Like, and I mean a simple one is like, I I bet you remember your first childhood phone number, right? And
44:45
Unidentified speaker (UNKNOWN): and
44:45
Unidentified speaker (speaker_1): then if you I've said sent to you now your business partner's cell phone number or somebody else like
44:50
Unidentified speaker (speaker_0): an MOY.
44:50
Unidentified speaker (UNKNOWN): MOY's
44:51
Unidentified speaker (speaker_0): cell phone numbers.
44:51
Unidentified speaker (speaker_1): Right. She
44:52
Unidentified speaker (speaker_0): doesn't
44:53
Unidentified speaker (UNKNOWN): like two
44:53
Unidentified speaker (speaker_0): minutes.
44:54
Unidentified speaker (UNKNOWN): Right.
44:54
Unidentified speaker (speaker_1): So like that's like a crazy thing, right? Like in the span of we're not particularly old, like I'm 46. I've gone from memorizing every phone number to I still have to now look up my sons, which is probably arguably the most important phone number I have because he's 13 and I'm traveling around Manhattan by himself. And if I lost my phone, I had to actually don't know how to get in touch with him right now because I
45:14
Unidentified speaker (speaker_1): never memorized it. But if you literally asked me, like
45:18
Unidentified speaker (UNKNOWN): my
45:18
Unidentified speaker (speaker_1): childhood phone number, my best friend's phone number, that's in there like forever. And so this book is really right. How do you bring back
45:26
Unidentified speaker (speaker_1): learning to readjust and bring your memory back into usage? And what are the skills and tactics that allow somebody to memorize not just one deck of cards in order, but 12 deck of cards in order? And so I just found it very interesting of uh the methodology that they use and trying to put that into practice a little bit more. So I'm not just tool reliant from an AI standpoint. I'm
45:46
Unidentified speaker (speaker_1): talking about I follow Ethan Moloch. I think he's really good about it. I then sort of every day have six or seven email newsletters that I get. I like Superhuman by Zane Khan, AI Fire, The Neuron.
45:59
Unidentified speaker (UNKNOWN): So
45:59
Unidentified speaker (speaker_1): I read those. I also have an AI sort of update and read. every one of them and provide me a daily synopsis based off of my role and what my key priorities are or what's what's being covered because you can't read everything. So I have it sort of it knows who I am, it reads the newsletters, like, oh, okay, this is you're the admin for our colloid account. They just announced a new feature around how to use
46:19
Unidentified speaker (speaker_1): memory more persistently. Great, I need to know that because I got to do a lunch and learn next Monday for our internal team. And I should tell them the new features that rolled out because they they're not going to spend time learning that. That's sort of on my job. So I got to talk to them and say, hey, memory is more persistent now. You can use it. Here's how it works. Here are the things to do, et cetera. Here's what you need to turn on in your own system.
46:39
Unidentified speaker (speaker_1): Right. So I do things like that. The last pieces, and I'll call this, which is you're on LinkedIn, I'm on LinkedIn. There's a lot of LinkedIn posts of like comment AI and I'll send you my 47 agent.
46:51
Unidentified speaker (speaker_1): thing. For the most part, it's a great lead gen tool, great for LinkedIn morality. Not always the most valuable.
47:01
Unidentified speaker (speaker_2): But
47:01
Unidentified speaker (speaker_1): I will now I will still take every one of those and I set up a time once a month
47:08
Unidentified speaker (UNKNOWN): where
47:08
Unidentified speaker (speaker_1): I go through all of the LinkedIn messages that get DM'd to me after I comment. I export all of them and then I base basically ask both ChatGPT and Claude to see whether or not they can recreate everything that's in there and give me an idea of whether or not this is real, valuable, or not. And that's I do that for marketing, anything
47:28
Unidentified speaker (speaker_1): consumer-related, B2B outreach. Because the truth is, I think
47:33
Unidentified speaker (UNKNOWN): in
47:33
Unidentified speaker (speaker_1): the same way that companies in Silicon Valley used to be garage-led, garage led, I now think there's so many people out there tinkering and building that you actually don't know who's going to put out a great piece on here's my prompt that did X, Y, and Z for me. And I don't have time to sort of immediately look at it, but I set aside one
47:53
Unidentified speaker (speaker_1): hour a month to go through, download, I put them into a folder. And then I literally just open that folder up in both tools and say, read everything that's in here. Tell me what's real or not real. And then we share it out. And when it works, we've built a a Slack community for all of our portfolio companies, for all the
48:13
Unidentified speaker (speaker_1): CEOs and anybody else that's sort of interested in AI. We invite them to Slack and we'll post sort of prompts, things that are cool. How do you optimize your Amazon listing?
48:23
Unidentified speaker (speaker_2): Okay.
48:24
Unidentified speaker (speaker_1): Somebody had a great idea about how to make your visibility better. I don't care where they're from or whether or not they have a PhD. If they've done something and it works, we should test it. And it's really easy to test now, so we should be doing that. And that's also allows us to find great speakers because we host a monthly AI office hours where we bring in outside speakers. And
48:44
Unidentified speaker (speaker_1): it's really one outside speaker and then two portcos portgos sharing something that they've done or built.
48:50
Unidentified speaker (speaker_0): Nice.
48:50
Unidentified speaker (UNKNOWN): Very
48:51
Unidentified speaker (speaker_2): nice.
48:51
Unidentified speaker (speaker_0): Appreciate that. If anybody wishes to get in touch with you to post this podcast, how best to do so, please?
48:56
Unidentified speaker (speaker_1): So happy to connect with everybody on LinkedIn. My email is public on our our websites. Our website's T Z P T is in Thomz and Zebra, P is in Paul Group.com. Right there you'll find my headshot bio, everything else has my full email. Happy to connect with folks. Always excited. to meet people that are
49:16
Unidentified speaker (speaker_1): building things in AI, entrepreneurs that have great companies that want help on thinking through capital, service providers that are building amazing tools or can be helpful.
49:30
Unidentified speaker (speaker_1): You need an army to build these things. Like it doesn't happen in a day, and you gotta have a lot of resources and you gotta provide the best people to solve everyday problems to folks. And so I'm happy to be people that are are great at solving problems.
49:43
Unidentified speaker (speaker_0): Sounds good. Well, thank you very much for coming on to the Private Activity Podcast, Jared. I've th thoroughly enjoyed this myself with regards to how you guys are leveraging, deploying, and adopting artificial intelligence. It's been phenomenal. Thank you very much.
49:55
Unidentified speaker (speaker_1): Thanks for having me, Alex. It was really fun.
49:57
Unidentified speaker (UNKNOWN): And
49:58
Unidentified speaker (speaker_0): as always, to all of our listeners, thank you for tuning in yet again to the Private Actu XT Podcast. Till the next time, keep smashing
50:04
Unidentified speaker (UNKNOWN): it.
This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.

More from The Private Equity Podcast, by Raw Selection

The Private Equity Podcast, by Raw Selection

The Middle Market Playbook: Discipline, Megatrends & Long-Term Value Creation

Episode discussion: Oak Hill Capital's Tyler Wolfram discusses disciplined middle-market investing and fiber infrastructure trends

Speakers’ claims: Tyler Wolfram of Oak Hill Capital outlines the firm's evolution from a family office to an institutional platform, emphasizing a disciplined investment philosophy focused on long-term value creation over short-term results. He highlights three core sectors: digital infrastructure, financial services, and essential services, using fiber-to-the-home as a primary case study. Wolfram argues that private equity's advantage lies in ignoring quarterly noise to manage for five-to-six-year outcomes. He notes that while LPs may forgive poor vintages, they do not forgive deviation from a stated investment mandate. The discussion concludes with an outlook on fiber market consolidation driven by strategic buyers like T-Mobile and AT&T, and the diverse exit routes available in the middle market.

Private EquityFiber InfrastructureMiddle Market
22 min0 assets10 claims to verify
The Private Equity Podcast, by Raw Selection

Underwrite the CEO Before the Company

Episode discussion: Ian Reynolds argues private equity firms should underwrite the CEO before the company to identify operational risks.

Speakers’ claims: Ian Reynolds, founder of Thesis Capital, critiques the siloed approach of large private equity firms, advocating instead for end-to-end deal ownership. He asserts that in the lower middle market, the management team is the primary differentiator among undifferentiated businesses. Reynolds emphasizes underwriting the CEO first, looking for high emotional intelligence, financial literacy, and adaptability. He suggests using operators in due diligence to identify skill gaps and targeting businesses capable of growing at twice the GDP rate.

Private EquityDue DiligenceLeadership
17 min0 assets10 claims to verify
All insights