Marty Milligan on AI governance risks, literacy mandates, and productivity gains for enterprises
Summary
Marty Milligan, founder of Digital On-Ramp WSI, discusses the shift from AI implementation to governance, citing burnout at EDS and the need for data protection. He identifies voice agents as a primary automation tool but warns of privacy risks in marketing platforms like HubSpot. Milligan argues that firms face legal liability for AI hallucinations, necessitating mandatory AI literacy similar to 1980s computer training. He claims AI adoption typically yields 30-70% productivity increases and plans to expand into home service businesses.
Key takeaways
- 1 Consumer-grade AI tools are discouraged for business use due to unpredictable data retention policies and risks of proprietary data being used for model training.
Unidentified speaker (speaker_2) · 11:19
Episode sections
- 0:33 – 4:33
Background and Career History
Marty Milligan discusses his 30-year career in government and enterprise IT, including managing a federal satellite weather data archive for NOAA. He describes leaving Electronic Data Systems (EDS) after its acquisition by HP, citing burnout and organizational changes as reasons for founding his own company in 2019.
- 4:33 – 6:33
Pivot to AI Adoption and Governance
Milligan explains how the COVID-19 pandemic prompted a shift from digital marketing to AI adoption. He notes that while many consultants focused on implementation, he bet on governance being the critical factor for client protection, leading to the formation of Digital On-Ramp WSI.
- 6:33 – 8:22
AI Adoption Trends and Voice Agents
The guest observes that ancillary tasks are the first to be automated across industries, with voice agents being particularly popular due to their cross-industry applicability. He highlights their utility in legal triaging, medical scheduling, and financial services, as well as their widespread use in marketing.
- 8:22 – 10:13
Governance Risks in Marketing Tools
Milligan warns that marketing departments often adopt AI-enabled tools like HubSpot without proper governance, leading to data privacy holes. He emphasizes the need for software inventories to identify AI plugins and establish policies on data usage and tool approval.
- 10:13 – 12:13
Top Governance Priorities
The speaker outlines three key governance priorities: approving specific tools, defining data rules for those tools, and establishing a fast review process to prevent employees from bypassing protocols. He advises against using consumer-grade technology due to unpredictable data retention and training policies.
- 12:13 – 17:43
Legal Liability and AI Literacy
Milligan cites legal sanctions resulting from AI hallucinations in law firms as a major governance failure. He compares current AI literacy needs to the computer literacy mandates of the 1980s, arguing that firms are liable for staff errors if employees are not properly trained to use AI tools competently.
- 17:43 – 24:48
AI Mastery and Productivity Gains
The guest defines AI mastery as moving beyond basic literacy to fundamentally changing business operations, similar to the web revolution. He shares a case study of building a database from PDFs in a weekend using ChatGPT and claims that AI adoption typically yields 30-70% productivity increases, often leading to hiring rather than job losses.
- 24:48 – 31:23
Future Focus on Home Service Businesses
Milligan states that Digital On-Ramp WSI plans to expand its governance client base and apply AI automation to home service businesses. The goal is to reduce back-office overhead, allowing small firms to scale revenue without linearly increasing administrative staff.
Claims worth verifying
Milligan spent 30 years in government and enterprise IT before founding his company in 2019.
Unidentified speaker (speaker_1) ·0:49
How to verify: Check Marty Milligan's LinkedIn profile or Digital On-Ramp WSI company biography for employment history and founding date. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Ross Perot founded Electronic Data Systems (EDS) in the 1960s and was paid approximately one billion dollars by GM to leave when GM acquired the company.
Unidentified speaker (speaker_2) ·2:08
How to verify: Verify Ross Perot's founding role and GM's acquisition of EDS via SEC filings (e.g., GM annual reports from the 1990s) or historical business records. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Primary sourceThe NOAA satellite data archive Milligan managed faced a situation where one satellite produced four times the data volume of the entire archive, with 16 satellites in the initial rollout.
Unidentified speaker (speaker_2) ·3:27
How to verify: Cross-reference NOAA's satellite launch schedules and data volume reports from the late 1990s/early 2000s with Milligan's tenure. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
HubSpot has integrated AI into its marketing tools, creating potential governance risks if users are unaware of data access policies.
Unidentified speaker (speaker_2) ·9:50
How to verify: Review HubSpot's official documentation or privacy policy regarding AI features and data usage for training. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
A law firm faced legal sanctions because an AI tool hallucinated non-existent cases, and the firm was held liable.
Unidentified speaker (speaker_2) ·13:15
How to verify: Search legal news databases or court records for sanctions involving AI hallucinations in legal briefs (e.g., the 2026 Matthew W. Schwartz case). For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Milligan built a database from a folder of PDF files using ChatGPT in one weekend.
Unidentified speaker (UNKNOWN) ·21:58
How to verify: This is a personal anecdote; verification would require Milligan's direct confirmation or a public demonstration/portfolio entry. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Primary sourceAI adoption typically yields a 30-70% productivity increase for businesses.
Unidentified speaker (speaker_2) ·23:49
How to verify: Compare this claim against industry reports from Gartner, McKinsey, or MIT Sloan Management Review on AI productivity metrics. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
The personal computer revolution took 15 years, and the web revolution took 12-13 years, whereas Milligan predicts the AI revolution will happen in 7-8 years.
Unidentified speaker (speaker_1) ·24:22
How to verify: Verify historical adoption curves for PCs and the Web against Milligan's forecast for AI, using historical technology adoption data. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.
Do your own research
- Verify the specific legal sanctions mentioned against the law firm using PACER or state bar association records.
- Check HubSpot's current data processing addendum and AI model training policies for enterprise customers.
- Review NOAA's historical data growth rates for satellite archives to validate the exponential growth claim.
- Compare reported 30-70% productivity gains against independent studies from Gartner or McKinsey on AI implementation outcomes.
- Assess the regulatory landscape for AI in home services, including state-level licensing and data privacy laws.
- Investigate Digital On-Ramp WSI's client list and case studies to verify the claimed productivity improvements.
- Check EDS's acquisition history by HP and subsequent organizational changes to confirm the timeline of Milligan's departure.
Full transcriptexpand
0:00Unidentified speaker (speaker_0): Welcome to the Entrepreneur Odyssey, the show where you learn from the best in class about how to become a successful entrepreneur.
0:09Unidentified speaker (speaker_1): Hello everybody, and welcome back to the Entrepreneur Odyssey. I'm your host, Techniat Selly, and this is the show where we sit down with founders who've been through the trenches of owning a business and they are generous enough to share their lessons with us. Before we dive into today's episode, a quick shout out to our sponsor, AI Efficacy, who helped their customers gain efficiencies and grow their businesses through smart and personalized workflows. Now, on to today's guest. Today's guest is Martin Milligan, who's the founder of Digital OnRamp WSI. Martin spent 30 years in government and enterprise IT, and at one point managing NOAA's federal satellite weather data archive. However, Burnout pushed him to found his own company back in 2019. When COVID hit, he made a bet that most consultants weren't making yet. And that was that AI adoption had a decade-long runway before
1:11it became commoditized, and that governance not just simply implementation, would be the thing that actually protected clients. He's since built a practice around exactly that, from various AI voice agents to various automations as well, all with the explicit goal of making his clients feel protected and protected. So we'll get into his founder stories and what he is seeing across industries. Martin, it's a a real pleasure to have you on the show today.
1:43Unidentified speaker (speaker_2): Thanks. I'm happy to be here.
1:44Unidentified speaker (speaker_1): Excellent. Let's start with the background. I know our our listeners would love to hear um hear the background. I mean, I mentioned it in in the intro there that you you know you've had some very prestigious roles. I'd love to hear about that moment where you decided, okay, you know what? I I I'm not gonna work for someone else anymore. I'm gonna take the reins myself.
2:08Unidentified speaker (speaker_2): It was a long time coming. I uh I worked for a little company called Electronic Data Systems. You may have heard a funny little guy with big ears from Texas ran for president one time. Ross Perough founded the company in uh the 60s. I I showed up right as they were showing Ross the door. Uh GM gave him about a billion dollars to leave. And by that time, he EDS was a huge conglomerate. I was played a major role in the insurance side of that, mostly the P and C, where we we would basically pilot advanced technologies in niche insurance markets. And if they panned out and we figured out how to implement them, then we would hand them over to other teams to roll out to bigger enterprises. Did that for a really long time, was very successful. Eventually, EDS was spun off from GM and then acquired by HP. And when HP acquires your large enterprise, it's a sign that they're about to fly it into the ground. So kind of, you know, mutually decided I should probably get out of there. And a friend of mine had a had a client looking for somebody to step in. They had uh they won a contract with NOAA to modern. modernize the data archiving of weather satellites. And the newest satellites that were look getting ready to launch, each satellite was four times the size of the data going into the the archive.
3:27So every day a s one satellite was producing four times what the archive was taking in, and there were gonna be, I think, sixteen 16 satellites in the first go-round. go round. So that was a an enormous upgrade. to the archive.
3:40Unidentified speaker (UNKNOWN): Uh
3:40Unidentified speaker (speaker_2): so they hired me to kind of come on and you know they kind of won the contract, but then it was like, well the good news is we won the contract. The bad news is we won the contract and now we have to build this. So they brought me in to manage that. It was very successful. We were under budget and on time. In fact we had a enough of a budget that we they I stayed on an extra month to put maintenance in place for the five year maintenance contract that followed on so they wouldn't have to pull the maintenance out of the new contract. We use the leftover money from the build out. At that point I was completely over doing anything with big business or the government. They had both pretty much worn me down to a nub. I really did not I I love the people. I love you know every you know people are great, but sometimes organizations can be really tough.
4:27Unidentified speaker (speaker_1): Yeah, that's bad to say.
4:29Unidentified speaker (speaker_2): And so I decided, you know, I'm gonna try and do something on my own. So I kind of I started, you know, doing my own software development and very quickly found out that I just didn't have the resources within West Virginia to to do what I wanted to do. And I was always scrambling trying to to to meet requirements. So I hooked up with a company called WSI, renamed the company, and we kind of expanded. from not just doing software development but also doing digital marketing and put together a nice runway of businesses that were coming on and then COVID hit. And what do you do? There we just wasn't enough to keep busy here. And while we tried to pivot some in-person businesses to online, it just wasn't really enough to satisfy what I wanted to do. And a lot of different folks in WSI were having the same experience. And we all shifted our attention to AI because that was something to keep us busy. We could learn about it. It looked like it was going to be the future. And then we started comparing notes. And then the WSI home office said, hey, you guys are you're changing the direction of the company, whether you know it or not. And so we could a huge pivot to doing a lot of AI adoption as a complete service. And now it's really a a a main focus of what we do is taking companies in and it doesn't have to be AI adoption as the as the the first engagement, but it's usually
5:48the goal of the of the complete engagement. So we may come in as a education or governance, we're just installing a voice agent, but the entire goal is to get the company to the point where they adopt AI and it becomes part of what they can do and it becomes part of their competencies as opposed to something that they are are still struggling with and need help with.
6:11Unidentified speaker (UNKNOWN): I
6:11Unidentified speaker (speaker_3): think that makes a lot of sense. And uh yeah, governance governance is is is an incredibly important area and you know something that we pride ourselves on um uh at your PCSI business I run uh Kiss me. Now you work with across a wide range of industries, for instance need more construction, energy, and even weather. And what's the trend in AI adoption you're seeing across all
6:39Unidentified speaker (speaker_1): of them? And where does it differ by industry?
6:42Unidentified speaker (speaker_2): Well, I I think you're starting to to see a commonality among all industries that ancillary tasks are some of the first things they often automate and some of the first things they concentrate on because they're not critical to the business and because they are things that are broad and cross cutting across industries, there's a lot of off the shelf and a lot of experience in doing that. So you'll see people adopt like the the voice agents are very popular because everybody can use a voice agent. I don't care what business you're in, whether it's legal, banking engineering, or you know, a a home service business, having a twenty four seven telephone answering service that can actually help people speak multiple languages, help make appointments, give some information about your services, answer some basic questions, collect some information is useful to everybody. So voice agents are that that that that's one technology that just cross cuts everybody. We like it in the legal field because we can do a lot of like triaging so that an appointment can be made with somebody who can actually help the person calling in quickly. We're probably going to be doing a little bit more branching out into the medical world on that one, because that seems to be a really difficult one. I think uh financial services are probably going to be something that's going to
8:02really benefit. from that in the near future.
8:05Unidentified speaker (speaker_1): Yeah, perfect.
8:07Unidentified speaker (speaker_2): Marketing's another another area where you're going to see just massive adoption.
8:11Unidentified speaker (speaker_1): Okay. Interesting. I I work with quite a few marketers and I'm seeing that. I'm seeing that trend. I'm also seeing that trend where marketers are implementing a lot themselves, which is not a bad thing. Um as long as they're doing it correctly. And to your point, keeping governance in mind.
8:31Unidentified speaker (speaker_2): Yeah, I think the the danger with the marketing side is that there's a little less you know marketing has always been kind of a a a a love-hate relationship with tech. And sometimes they do a really good job adopting tech, and other times they just kinda kind of they don't really have the background to know. So we like to we really like the idea of adopting AI into to marketing technology. We do like to caution people to do some governance first. Because what we're seeing is a lot of these AI tool or a lot of these existing marketing tools are adopting AI plugins and AI capabilities.
9:08Unidentified speaker (UNKNOWN): Yes. And
9:08Unidentified speaker (speaker_2): so a lot of companies are getting AI enabled tools in their marketing, but they have haven't actually put governance in place on how to use them and what data is allowed to be used into them. So the caution with marketing is that it's it's a place that is really it it's been very fast to adopt the technology.
9:29Unidentified speaker (UNKNOWN): Yeah.
9:30Unidentified speaker (speaker_2): But a lot of the technology that they've adopted is more of a side effect than an an an actual decision. So they don't necessarily know. Like, for example, we sell HubSpot. HubSpot's a fantastic tool. It has integrated AI into a lot of what it does, but if you're not aware that HubSpot's using AI and what data it has access to, you might have a governance hole there that you're not really thinking about. So we really like to to sit down with companies that are doing these these technology adoptions. as a side effect and make sure that they're incorporated into their governance. So part of our governance part of our governance tasks is to do a software inventory and a tool inventory to figure out, you know, maybe you're having AI that you don't even know about and you might need a policy about how to use it and what what data is available to
10:21Unidentified speaker (UNKNOWN): it.
10:22Unidentified speaker (speaker_1): Nice. Yeah, that that that makes sense. And when a marketing agent agency, an insurance agency, whatever, whatever industry they're in, are thinking about governance, what's the first thing or what's the top three things, let's say, they need to be thinking about?
10:39Unidentified speaker (speaker_2): We'd really like the idea of thinking about the what tools you're going to approve and what data you're going to allow into those tools. And lastly, because so many tools are being rolled. out so quickly, if somebody needs something, how can they get that reviewed and approved quickly so that they don't try to get around your processes? So, you know, the these those are those are really management decisions, not necessarily AI decisions. And so having somebody help you make an informed, you know, policy about what can be used. For example, you really don't want to be using any of the entry entry-level levels. consumer grade technology because as you know, they have different, wildly different data retention policies. Some of your data can be used in training, it can be used in search. So obviously the the consumer grade stuff is is out of the question. It's really what level of professional grade grades you need. And you know, are you going to set up a teams team and how how do you share data among people, among your team members? and that kind of thing. So it's really those decisions at that level that we like to see made before the team starts going wild. I mean, we know there's going to be a lot of people on the floor who are going to embrace technology. That's just the way this works. And it's best to be able to make those people enabled because
11:59there's a discrete policy and they can work within the bounds and they don't feel constrained as opposed to having them do everything on the sly. So that's you know making sure you get out ahead of that and have those decisions made and you're you know fully capable of that. The other so that's that's that's really the the the top three things before you start educating and all that other stuff.
12:22Unidentified speaker (speaker_1): Sure. Okay. That that that makes sense. Let me ask you a flip question to that, if I may. Um I I know you work with a lot of legal firms or law firms, and you've seen cases where AI use led to actual legal sanctions. What's the biggest governance mistake you see clients make before they come to you and and how do you specifically fix it?
12:47Unidentified speaker (speaker_2): It's the the the assumptions about what the tools do is probably the biggest mistake.
12:52Unidentified speaker (speaker_1): Okay.
12:53Unidentified speaker (speaker_2): Um you know the idea that if you just pick the right tool, it's gonna be solve your problem.
12:59Unidentified speaker (UNKNOWN): That
13:00Unidentified speaker (speaker_2): you know picking the right tool is obviously important. You don't want to pick the wrong tool.
13:04Unidentified speaker (UNKNOWN): On
13:04Unidentified speaker (speaker_2): the other hand, you can pick the right tool and have have still have problems. So
13:10Unidentified speaker (UNKNOWN): in
13:10Unidentified speaker (speaker_2): fact we did a a YouTube video on a uh an unfortunate law firm that actually used a legal specific tool, but it also hallucinated. And you're an AI, you're gonna have hallucinations. You
13:24Unidentified speaker (UNKNOWN): really have
13:25Unidentified speaker (speaker_2): to understand that the lawyer is the lawyer. The tool is the research
13:31Unidentified speaker (UNKNOWN): assistant.
13:31Unidentified speaker (speaker_2): If you assign a task to a research assistant,
13:34Unidentified speaker (speaker_1): yeah,
13:35Unidentified speaker (speaker_2): and they come back with a case that doesn't exist or they misunderstand something, and you use that in your brief, you are still responsible for that. And AI doesn't change that. So the idea that
13:49Unidentified speaker (UNKNOWN): if
13:49Unidentified speaker (speaker_2): you just pick the right tool, you don't set bounds on how to use it, or you know, every reference that it makes has to be researched, every statement that it or every fact that it insists is true, you have to verify. And the important, more important one is that if you bring a tool into your business, especially in the legal world, you have an obligation to make sure that your staff is competent to use it. And if you're not, that that becomes a liability issue The for you. And this is something that has been a problem with technological adoption among the professionals for a really long time and hasn't been acknowledged and hasn't really been addressed. Yeah, I can remember back in the late 80s when computer literacy was all the rage, and now we're dealing with AI literacy.
14:41Unidentified speaker (UNKNOWN): Yes.
14:41Unidentified speaker (speaker_2): And
14:41Unidentified speaker (UNKNOWN): there were
14:42Unidentified speaker (speaker_2): a lot of problems back then where you just had people who were not competent in the tools they were using making big mistakes and having companies on the hook for a lot of the errors being made. Today we're starting to see the same thing where you'll do everything right. You'll you'll pick out a tool that's specific to your industry. You'll restrict the data that goes into it. You'll you know have like a that that quick training class, but then you won't train your paralegals or you won't train your secretaries or your your your executive assistants and somebody will misuse a tool and the person who is carrying out a function is not competent for the tool, and then all of a sudden you create a liability for yourself. So making sure that you, you know, and that's really where we're trying to make a difference is making sure the companies understand you've got this massively powerful tool and you're going gonna to be putting it in the hands of your it's like, well, you wouldn't put a machine gun in the hands of your eight year old.
15:43Well, you know, I'm in West Virginia, so uh but you know, the idea is that you don't wanna give, you know, everybody this massively powerful tool and have them not be able to control it or understand how to use it. And like I said, I you know, I came of age back in the eighties 80s when the personal computer was replacing the mainframes for a lot of office tasks. And it was, you know, for larger companies, it was, it was completely like just that the water you swam in was educating on computer literacy. Everybody from the person answering the phone to your senior consultant engineer engineers were all going through that computer literacy learning curve.
16:25Unidentified speaker (UNKNOWN): And then
16:26Unidentified speaker (speaker_2): they had to figure out what tools they were going to use and go through the mastery curve on those. And all of a sudden there was this enormous boom in productivity in the 90s because we went through with that in the late 80s. And we're getting ready to do exactly the same thing. We've got this massively powerful tool. People are kind of realizing that and they're having to go through the AI literacy phase. And then they have to finish figure out what they're going to be using for their business and get the mastery of those tools. And so trying to get people we have been through this, but we for we forget, right? Generation to generation. Yeah. The
17:02Unidentified speaker (UNKNOWN): people
17:02Unidentified speaker (speaker_2): who are running businesses in the eighties are all retired. The people who are running them now, they didn't by the time they came onto the scene, computer literacy wasn't something they had to worry about.
17:12Unidentified speaker (speaker_0): Right.
17:13Unidentified speaker (speaker_2): So they uh they this this idea that now we have to get everybody AI literate and then we have to get them AI mastery is probably something they hadn't thought about until they started actually trying to do it in their business.
17:26Unidentified speaker (speaker_1): Right. Absolutely. Couldn't agree with you with you more. It is a transition. And we uh, you know, I think the the transition will probably be faster than the personal computer. transition, just by the nature of how fast AI is progressing every single day and week. But uh yeah, it's it's definitely an interesting time that we're living through as business owners for sure. So I think I mentioned it earlier. Your philosophy is training clients to eventually not need you and handing over AI capability rather than creating. dependency, which is unusual an unusual stance for a consultancy, but a very admirable one at that. Why build the business in that way?
18:13Unidentified speaker (speaker_2): Well, because we think that the we think that the market is gonna move beyond where it is today. And uh people are very quickly going to develop AI literacy, and that the you know the mass massive benefit that we're seeing with AI literacy is going to pale in comparison to the benefit we're going to see with AI mastery. And that we really need businesses to get to the point where they can do their own AI literacy. They're an AI literate company, and that they need to move into the mastery phase. And the mastery phase phases is where we're really going to see a a large takeoff. And if we have to kind of share shackle ourselves to the AI literacy part of the curve, that's really going to slow down how fast we can help our clients. So our goal really is to get our clients' businesses into the AI through the AI adoption curve, get them through AI literacy, and have them start working on doing things that that move them into the AI mastery, which is, you know, have the have governance. built into all parts of their organization, have you know their their basic AI automations addressed and have them start changing the way they do business to actually be AI native so that they they don't it's not just we're gonna take these things we used to do and automate them. Now it's going to
19:32be we've got this new way of doing business, let's automate that. And so I think that the the automation is the AI automation space and the productivity space is really going to explode in probably the next 10 years, where businesses are going to, they're not just going to adopt new technology. They're going to change the way they did everything. It's going to be like the late 90s and early 2000s when businesses didn't just automate their existing workflows with computers. They completely changed how they did. business fundamentally. You know, the web, per personal computer, client server, then to the web, completely changed the way business got done. And we think that's going to happen with AI, only it's going to happen in a very compressed timeframe.
20:17Unidentified speaker (speaker_1): Yeah.
20:17Unidentified speaker (speaker_2): So our goal is to get all of our clients' businesses through that AI adoption, get them AI literate, have them focus on AI. mastery where their core business functionality is concerned, and have them move into native AI where they don't need us for the day today. They need us for those advanced topics that are really going to take them to the next level.
20:42Unidentified speaker (speaker_1): That makes perfect sense. And you've mentioned obviously AI mastery. For our listeners, what in your words do you mean by AI master? mastery? Because we know we're all using uh our various LLMs. Some of us may even be creating skills or agents, but what do you define as AI mastery?
21:01Unidentified speaker (speaker_2): And I think the uh the the place where we're gonna start seeing AI mastery first is probably in the software development end of things. And that's probably where the that where the the the way we work with computers on a fundamental level is changing. with software development today. You can see it
21:18Unidentified speaker (UNKNOWN): you
21:18Unidentified speaker (speaker_2): know the the we we went through vibe coding. Now we've got agent coding. And I don't know if you've played with some of the stuff. I mean, my eye opener was I had a startup come to
21:29Unidentified speaker (UNKNOWN): me
21:29Unidentified speaker (speaker_2): and a friend of mine was really trying to get this thing off the ground. And he needed a database as part of building this. And all he had was about a giant folder full of PDF files. And it was what do you
21:43Unidentified speaker (UNKNOWN): do with
21:43Unidentified speaker (speaker_2): this, right? And so we tried everything. We tried writing normal code. We tried hiring you know VAs to look at it and you know do scans and build the database that way. And in a weekend I sat down with ChatGPT and built the database.
21:58Unidentified speaker (UNKNOWN): Now
21:58Unidentified speaker (speaker_2): we still had to do QA on it,
22:01Unidentified speaker (UNKNOWN): but
22:01Unidentified speaker (speaker_2): I was able to do the entire thing in a weekend with Chat GPT. I couldn't even run I'm a um I'm I'm no longer like the a great programmer, but there was a time when I would have I was a I would have put my slinging code abilities with anybody.
22:16Unidentified speaker (speaker_1): Right.
22:16Unidentified speaker (speaker_2): Today I'm a little rusty and I don't do it that but you move into management. You you you have that management lobotomy and you don't code that well anymore. But you know I'm like I said I'm I consider myself you know, pretty capable when it comes to to doing code. And I, you know, we'd all taken a crack at this and we could not build it. ChatGPT did it in a weekend. And it wasn't that hard. And it was an eye opener that look, there are things that we used to be able to do programming wise that were just massive time sinks. Some of them were impossible. And now all of a sudden we have tools that can make this happen. in, you know, in a trivial amount of time. It's like you're seeing all this data center build out because the cost of software is dropping so bad. So when we talk about AI mastery, if you look at what's happening on software development, that is a good example of an industry that was, you know, really advanced, really pushing hard, and all of a sudden had this disruptive tool that just blew blow
23:14Unidentified speaker (UNKNOWN): up.
23:14Unidentified speaker (speaker_2): every has blown everybody out of the water. We're doing things that we never thought were possible. You know, a a team of three or four people can do things that used to take fifty, sixty. That's where AI mastery is going to all parts of the business. You know, today,
23:28Unidentified speaker (UNKNOWN): I
23:29Unidentified speaker (speaker_2): don't know what you guys see on productivity, but when we come into a business and start doing automation, you might get a 30 to 70 percent productivity income. increase, but you're not completely, you know, this idea that AI is going to come in and take everybody's job is just not there. Usually what happens is you do a 30 to 70% improvement, and all of a sudden there's things that can get done, and they're looking to hire people, not get rid of them.
23:55Unidentified speaker (speaker_1): Yeah.
23:56Unidentified speaker (speaker_2): When you start getting into mastery, you're not just having that 30 to 70 improvement, you're building out entire capabilities and businesses and parts of the business that didn't exist before,
24:08Unidentified speaker (speaker_1): and
24:08Unidentified speaker (UNKNOWN): all of
24:08Unidentified speaker (speaker_2): a sudden new products and services are coming to market quicker. I I think that that that revolution is something that people just haven't gotten their mind wrapped around because you know, we had 15 years with the PC revolution.
24:22Unidentified speaker (speaker_1): Yeah.
24:22Unidentified speaker (speaker_2): We had, you know, 12, 13 years with the web. This is gonna happen within seven or eight years. And we, you know, it's it's just something we have to get ready for because it's gonna happen.
24:34Unidentified speaker (speaker_1): I've heard a lot less than that, even just because of the speed that that we're we're um developing this technology and what you can do with it. But yeah, I I think very well said. Thank you. So I'd love to hear what is next for digital on-ramp uh WSI over the next sort of six, twelve months. Are you working on more governance work? What's the the future looking like?
24:59Unidentified speaker (speaker_2): We've got mostly we really are gonna focus on on growing our our client base on the governance side. Um I have some very specific software development projects that I can't really talk about. Um we've one of the things when I started this, we were doing a lot of assistance with uh startups. And some of those are they're gonna change directions. And I don't with a startup you never know. Um you know one of the guys we work with ended up collapsing the company because the the people were more valuable than the idea.
25:40Unidentified speaker (speaker_1): Interesting.
25:41Unidentified speaker (speaker_2): So a lot of you know a lot of the companies we've been working with disappear that way. They just become it's like, oh you've you've accomplished this, great. We love the tech, but we love the fact that you guys went through the learning curve first.
25:52Unidentified speaker (speaker_1): And
25:53Unidentified speaker (speaker_2): we want
25:53Unidentified speaker (UNKNOWN): that
25:53Unidentified speaker (speaker_2): learning curve experience. The tech, yeah But my it's a nice side idea, but it's not really gonna sell. So
26:01Unidentified speaker (speaker_1): okay, interesting.
26:02Unidentified speaker (UNKNOWN): So
26:02Unidentified speaker (speaker_2): we're probably going to see some some more capabilities come out on being able to to to build a company with small with a smaller number a more functional company with a smaller number of people is probably the is the direction we're going.
26:18Unidentified speaker (UNKNOWN): Okay.
26:19Unidentified speaker (speaker_1): Okay. And that makes sense. And how large is the company at the moment then in terms of number of people?
26:25Unidentified speaker (speaker_2): WSI it depends on how you count WSI. It could be it could you know, think of in the the in the hundreds, but you know, each office is only a few people. But we're looking at taking the ability like for example, you can run a a three person office and get a lot done, but in the home service industry, it's really hard to be a you know, a three person company. of any size because you're so what we're really looking at doing is can we take like this the way we would grow the productivity of a service is of a a purely services business and bring that into the real world. So what can we if you own a home service business,
27:08Unidentified speaker (speaker_1): how
27:09Unidentified speaker (speaker_2): much overhead can we remove from it so that you don't have a lot of excess employees who are not, you know, the the the the the the warhead on forehead type of guys?
27:22Unidentified speaker (UNKNOWN): The
27:22Unidentified speaker (speaker_2): employees who are out there actually that have to interface with the client, do something for the client. Those guys, you can make them a little more productive, but we could what can we take out of the business that you don't have to pay for in order to grow a home services business so you can concentrate on a accomplishing client services instead of back office services and all that kind of stuff. So that that's going to be the the concentration is going to be saying, you know, it's easy to go to a a law firm
27:50Unidentified speaker (speaker_1): and
27:51Unidentified speaker (speaker_2): make them more productive. It's easy to go into a big accounting firm and make them more productive. Can you do the same thing for somebody who's like a construction company or you know those kind of things so that they don't have a lot of back office that they need to grow. So if
28:06Unidentified speaker (UNKNOWN): they
28:06Unidentified speaker (speaker_2): want to go from a million dollars a year to ten million dollars a year as a home service business,
28:11Unidentified speaker (speaker_1): can
28:12Unidentified speaker (speaker_2): you do it without adding a whole bunch of overhead on the back end?
28:15Unidentified speaker (speaker_1): I see. Well, that's a big challenge. That's an interesting challenge because you you need uh for every construction uh uh project, you need people on the ground, right? So
28:26Unidentified speaker (UNKNOWN): But you're talking
28:27Unidentified speaker (speaker_1): more back a uh back office function.
28:29Unidentified speaker (speaker_2): When I go into to businesses in West Virginia and I ask them, you know, do you need marketing services? And almost all of them have the same answer. It's like, you know, before I need marketing services, I need people. There's plenty of of effort being put into the training of people who actually go out and do the work. And now our question is, how much of that back back-end end support can we basically remove with AI agents so that they're not necessarily a linear growth?
28:59Unidentified speaker (speaker_1): Yeah, makes sense. Makes sense. Gosh, we're almost at time, Martin. So I just I want to ask you one more question. If a small business owner is listening today, has never really touched AI or you know maybe just use it as a chat bot or something like that and doesn't necessarily know where to start. What's the first step you'd tell them to take this week?
29:20Unidentified speaker (speaker_2): One of them is we have a uh a Chat GPT introductory document that really does a good job.
29:26Unidentified speaker (UNKNOWN): If you go
29:26Unidentified speaker (speaker_2): through that, if you download that document and go through it, you're gonna have to be way ahead of your knowledge on what it can do for your business. What's with some ideas about what other business owners have tried to do, what challenges. they've faced, how they've overcome them. You can read that survey. We're right now conducting the third year of that survey. An assessment you can take to figure out how how close your business is to being ready to adopt AI and how much how much more work you're going to need to do to make to actually have it have AI do something in your business. It just basically gives you some idea of how prepared you are and some of the things you might want to do in order to to start taking advantage of it fear. Or double your money back. So they
30:10Unidentified speaker (UNKNOWN): can get a a free
30:12Unidentified speaker (speaker_2): hour to come in and talk and we can talk
30:14Unidentified speaker (UNKNOWN): about
30:14Unidentified speaker (speaker_2): where they are and what they want to do. And we could talk about, you know, just how ready are they,
30:18Unidentified speaker (UNKNOWN): you
30:19Unidentified speaker (speaker_2): know, where what what dangers they face and that kind of stuff, somebody.
30:23Unidentified speaker (speaker_1): Nice. Nice. Excellent. Well, Martin thank you so much for coming on and being so uh open about the journey and also giving a ton of great advice to our listeners. Really, really appreciate it. And um, to my listeners, if you enjoyed today's conversation, go check out Martin and the team at DigitalonRamp Digital on ramp WSI. You'll find them at digitalonramp. digital on ramp w of course I will add uh the link to the show notes as always. And you can also reach Martin on LinkedIn and follow DigitalonRamp WSI on the LinkedIn page as well. Martin, uh I hope you've enjoyed this as much as as much as much as I have today. Thank you so much. Well,
31:08Unidentified speaker (speaker_2): thank you. You're a great host. I appreciate it.
31:10Unidentified speaker (speaker_1): Thank you. And thanks to all the listeners. Please do uh follow the podcast and subscribe so you don't miss out on any other episodes. And until next time, keep building.