JP Morgan's Camila Penna discusses how AI is shifting content consumption from human reading to machine interrogation.
Summary
Camila Penna, co-head of Global Cash Equity Sales at JP Morgan, explains that while demand for original information remains constant, AI is fundamentally altering how investors consume it. She describes the shift from traditional human reading to 'machine interrogation' via Large Language Models (LLMs). Penna details JP Morgan's 'Content Hub,' which centralizes over 100,000 pieces of annual content and 20,000 sales emails for programmatic querying. She notes that while only about 10% of clients currently use these feeds, the technology automates low-value tasks, allowing salespeople to focus on high-value dialogue. The discussion highlights risks around intellectual property, trust, and the high cost of AI infrastructure.
Key takeaways
- 1 JP Morgan has an industry-leading research franchise with around 500 authors publishing over 100,000 pieces of written content annually.
Camila Penna · 3:41
- 2 The cash equities sales organization has over 400 people writing over 20,000 emails a year across 25 different locations.
Camila Penna · 4:01
- 3 Clients are using content feeds for signal generation, research synthesis, thesis monitoring, and idea discovery.
Camila Penna · 7:11
Episode sections
- 0:04 – 1:25
Introduction and AI's Impact on Consumption
Edwina Lowe introduces Camila Penna, who notes that AI is shifting content consumption from human reading to machine interrogation.
- 1:25 – 3:15
Content as Data and Machine-Readable Formats
Penna explains that LLMs have increased demand for machine-readable content beyond traditional quant funds.
- 3:15 – 5:23
JP Morgan's Content Hub and Workflow Changes
JP Morgan created a 'Content Hub' to centralize 20,000+ annual sales emails for AI-based querying.
- 5:23 – 6:48
Salesperson Role and Efficiency Gains
Centralization automates low-value tasks, freeing salespeople to focus on original insight and client dialogue.
- 6:48 – 8:18
Client Usage and Signal Generation
Clients use feeds for search, summarization, and signal generation, with some integrating content into risk hedging workflows.
- 8:18 – 10:02
Client Sophistication and Infrastructure Costs
Only about 10% of clients consume these feeds due to high infrastructure costs and varying sophistication levels.
- 10:02 – 11:55
Trust, IP, and Commercial Value
Commercial value lies in enterprise rights to consume proprietary IP with proper governance and attribution.
- 11:55 – 18:01
Risks and Future Innovation
Risks include disintermediating creators; future innovations include expanding to macro/commodities and searchable conference transcripts.
Claims worth verifying
AI is changing the consumption model of content, shifting from human reading to machine interrogation.
Camila Penna ·1:23
How to verify: N/A (Claim about market trend/behavior described by speaker) For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Quant funds have consumed written research through machine-readable formats for over 10 years.
Camila Penna ·2:13
How to verify: Verify via JP Morgan investor relations or press releases detailing the history of their machine-readable research distribution. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Primary sourceJP Morgan has an industry-leading research franchise with around 500 authors publishing over 100,000 pieces of written content annually.
Camila Penna ·3:41
How to verify: Verify via JP Morgan investor relations or press releases detailing the scale of their research franchise. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
The cash equities sales organization has over 400 people writing over 20,000 emails a year across 25 different locations.
Camila Penna ·4:01
How to verify: Verify via JP Morgan investor relations or press releases detailing the scale of their sales organization. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Only about 10% of JP Morgan's thousands of clients are currently consuming these content feeds.
Camila Penna ·9:49
How to verify: Verify via JP Morgan investor relations or press releases detailing the adoption rates of their content feeds. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
JP Morgan is expanding Content Hub to macro, commodities, and credits.
Camila Penna ·15:04
How to verify: Verify via JP Morgan investor relations or press releases detailing the expansion of their Content Hub. 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 JP Morgan's annual report or investor presentations for the stated number of research authors (500) and content volume (100,000 pieces).
- Check JP Morgan's public disclosures or press releases regarding the 'Content Hub' launch and its specific features for enterprise clients.
- Review regulatory filings or compliance statements from JP Morgan regarding cross-border data governance and intellectual property rights for AI training.
- Assess the cost structure of AI infrastructure by reviewing industry reports on LLM deployment expenses for institutional clients.
- Look for independent case studies or white papers from other asset managers describing their adoption rates of AI-driven content feeds.
- Investigate the specific APIs or protocols (e.g., MCP) mentioned for data delivery to understand technical integration requirements.
- Search for competitor offerings from other bulge bracket banks (e.g., Goldman Sachs, Morgan Stanley) to compare content aggregation strategies.
- Review academic or industry literature on the efficacy of LLMs in generating alpha versus merely summarizing existing information.
Full transcriptexpand
- 0:04
- Edwina Lowe: Hi, and welcome to JP Morgan's Making Sense. I'm Edwina Lowe from our Data Assets and Alpha Group, and today I'm delighted to be joined by Camilla Penner, who is co-head of Global Cash Equity Sales. So, Camilla, thank you so much for joining me on the podcast today.
- 0:18
- Camila Penna: Thank you for having me, Edwina.
- 0:20
- Unidentified speaker (UNKNOWN): So,
- 0:20
- Edwina Lowe: Camilla, could you start by introducing yourself and your role here?
- 0:24
- Camila Penna: Yeah, so I am co-head of Global Cash. Equity Sales at JP Morgan. I've spent more than two decades in financial markets across New York, Brazil, and London. And I've always been extremely interested just in technology, data, AI, how it's been reshaping the way that investors make decisions. When I think about when I
- 0:44
- Camila Penna: started back in the financial markets, I remember waiting on faxes. Then when BlackBerries arrived, it felt completely transformational. Then we started to see changes in electronic trading platforms, data platforms, and just the pace of technology changing accelerated enormously. But the reality is,
- 1:04
- Camila Penna: I haven't seen a pace of transformational change like what we are going through today. Investment managers still want original information, differentiated thinking. market intelligence, good ideas from salespeople. None of that has ultimately changed. However, AI is changing the
- 1:23
- Camila Penna: consumption model.
- 1:25
- Edwina Lowe: Yes, I think it's fair to say that what we're seeing is content, which in and of itself has not changed, becoming a form of data in and of itself.
- 1:34
- Camila Penna: Yes, so reality is that written research remains the most prominent product. and the one originating the most market dialogue. But content is much broader than traditional research. It's also a salesperson's morning note. It's themes from a company meeting. It comes from our conferences, from spec sales commentary,
- 1:54
- Camila Penna: from market intelligence, from our traders providing color on flows, volumes, positioning, analyst discussions. So content is actually extremely broad and becoming broader. So for over 10 years, we've had quant funds that have consumed our written research through machine readable
- 2:13
- Camila Penna: format. Machine readable research, in other words, content as data, is not necessarily a new thing. It wasn't until LLMs were introduced that the demand started to trend drastically. So this is no longer exclusively about a quant consuming a data set. PMs, analysts, fundamental firms
- 2:33
- Camila Penna: are now wanting to interrogate content. We are seeing investment teams that build agents around these contents. We have traders that want to consume sales and trading color programmatically, Dramatic internal platforms that can combine content with proprietary and third party. third-party data. We're seeing so many different
- 2:53
- Camila Penna: usages of content as data now on the street that the reality is that this is what is really transformational around the content space.
- 3:03
- Edwina Lowe: That really begs the question: what are we doing about this here at JP Morgan? And in your seat, Camilla, I know you spend a lot of time talking to clients about what they need and what they're looking for from our markets business.
- 3:15
- Edwina Lowe: How have you thought about implementing that change into our workflows at JP Morgan?
- 3:21
- Unidentified speaker (UNKNOWN): So,
- 3:21
- Camila Penna: Edwina, you're right. This content journey is a perfect example of where clients have made us better. Their expectations of where the content world is going has forced us to rethink our own workflows and our own investments. around content. We have
- 3:41
- Camila Penna: an industry-leading research franchise that has around 500 authors publishing over 100,000 pieces of written content annually. And it was already way advanced in the journey of machine-readable distribution. As I mentioned, a lot of the quant funds were already consuming it. What else could we be applying the same principles of how we were delivering the
- 4:01
- Camila Penna: machine-readable machine readable research to other? contexts across the market sales organization? To use an example of the cash equities world, salespeople across our organization create content every day. And just to give you an idea of our scale, we have over 400 people around the world writing over 20,000 emails a year. We have people in 25
- 4:20
- Camila Penna: different locations, and the reality is it was just too fragmented. So we created content hub, essentially. it's a one-stop shop for market commentary. It centralizes the content written across our sales organization. I'll give you like a practical example. We come in in the morning in London and Asia already is in half
- 4:40
- Camila Penna: their day. There's a lot of salespeople that have already written content throughout the day. And now anyone can just go into Content Hub and see all of these pieces. the that have been written throughout the morning. You are able to query it through our AI search engine and prompt it to be
- 5:00
- Camila Penna: able to already deliver all of the key insights that one should know of all of the aggregated content that was written in that morning. That's something new that we've never had in the past. And it wasn't until we were able. to aggregate all of the content into a single place, into a single depository, that
- 5:20
- Camila Penna: we're able to find uses like that.
- 5:23
- Edwina Lowe: Absolutely. And really, my natural question would be to ask how that impacts the day job of a salesperson and of course their client relationships.
- 5:33
- Camila Penna: So there was a lot of questioning when we created Content Hub around does this make the salesperson actually less important? And funny enough, I actually think that it's the exact opposite. Technology ends up automating low-value tasks, copying, forwarding, reformatting, searching, redistributing,
- 5:53
- Camila Penna: and multiple rewriting of the same material. That arguably is what goes away with centralizing the content into a single place. This creates actually more time for original insight to be created, client dialogue, ideas, interpretation, connecting information across the platform, and ultimately
- 6:13
- Camila Penna: just having more time to have proper dialogue discussions and see clients face to face. Relationships have always been and continue to be at the heart of what we do, it's the most important thing in the sales process. salesperson's day-to-day.
- 6:29
- Edwina Lowe: The efficiency gains are loud and clear. So you're making this content coming from the markets business on top of the research machine readable content that we've done for a decade or so available via API or MCP. What are clients actually doing with these fees feeds and how is it
- 6:48
- Edwina Lowe: informing their own investment workflows?
- 6:51
- Camila Penna: So the main usage, I would say, really is to be able to put this content as data into LLMs, to be able to search, summarize, retrieve relevant materials, and just being able to prompt and surface information faster, quicker. We've seen a few clients that have been able to
- 7:11
- Camila Penna: take this to the next level and start to find. signal generation, research synthesis, more thesis monitoring, idea discovery. Some have been able to insert this into their risk hedging process. And we're starting to see clients creating agents that feed relevant
- 7:31
- Camila Penna: commentary into already their existing workflows. Does this actually change the relationship between trading and investing? And the reality is that a trading desk equipped with better information can become more consultative to the investment organization, bringing market context, flow, positioning, related
- 7:51
- Camila Penna: ideas, and risk indicators back to PMs, it's moving useful information closer to the point of decision making.
- 7:58
- Edwina Lowe: So we've discussed how, up until fairly recently, it really was the quant funds who were the primary consumers of systematic data. But arguably, as you've said, the rise of LLMs has really democratized and leveled the playing field when it comes to the consumption of data across the full breadth of
- 8:18
- Edwina Lowe: client types and strategy types. What's your take on this? How sophisticated are clients today when it comes to evolving. their use of these types of tools?
- 8:30
- Unidentified speaker (UNKNOWN): So,
- 8:30
- Camila Penna: in theory, Edwina, that is true. However, the dispersion is still vast. There's still quite a big spectrum here of the differences within the clients and what they want. So, on the one end, you have some clients that are quite advanced in this sophisticated data and AI infrastructure build. They're coming into us with new requests of
- 8:50
- Camila Penna: unstructured. content and really pushing us to think about what are the arts of the possible. And then you have another end of clients that are just simply still in the journey of consuming machine readable research through API feeds, and they're really just looking for efficiency gains for their PMs and their analysts and are really
- 9:09
- Camila Penna: using LLMs to swift through larger amounts of content as data. You know, AI is still very expensive. Building out your new technology or new infrastructure requires a huge lift. And I would say that for many of them, it is still the very beginning of this journey. And so we
- 9:29
- Camila Penna: are here to service all styles of clients, and we are here to service our clients throughout where they are in the journey, not necessarily where we are in the journey. So, for example, we have thousands of clients that we service. And I would say it's only 10% of them that are probably consuming these feeds today. So,
- 9:49
- Camila Penna: as you can see, there is still a wide journey to go through.
- 9:54
- Edwina Lowe: And what does that mean when it comes to content availability, content explosion, a world in which there is infinite content?
- 10:02
- Unidentified speaker (UNKNOWN): Yeah.
- 10:02
- Camila Penna: look, the amount of available content is exploding at a pace that I think no one imagined possible. Now, more information does mean more noise, which increases the premium on trust and differentiation. So from JP Morgan perspective, there's so much rigor in
- 10:22
- Camila Penna: everything that we do, be it from creation of content. from making sure that the delivery mechanisms meet all of the right standards. We go through compliance, we look at cross-border regulations. I would say that there should always be an extra confidence from the client around what it is that we're
- 10:42
- Camila Penna: delivering, how and why. And so this isn't really about commercializing and monetizing an API. We're also recognizing the enterprise rights to a premium intellectual property as data that carries a lot of value.
- 10:58
- Edwina Lowe: It's really clear to me that it's imperative that the end consumer needs to have absolute confidence in the credibility of the content that they receive. And can you expand upon that point on commercial value?
- 11:15
- Camila Penna: So we invest heavily in differentiated intellectual property and in the people who create it. The value that we see isn't simply in a delivery mechanism. It's in the right to consume the content programmatically, systematically across an enterprise with structured metadata, inside proprietary applications, an AI
- 11:35
- Camila Penna: workflow, with appropriate governance, appropriate attributions. attribution. And the reality is that the arts of the possible of delivering content as data is so drastic that we see a different value today with the possibilities that AI brings than we ever have in the past.
- 11:55
- Unidentified speaker (UNKNOWN): So,
- 11:55
- Edwina Lowe: Camilla, I it would be remiss of me not to ask about the risks. What could go wrong? This is all very new.
- 12:04
- Unidentified speaker (UNKNOWN): I
- 12:04
- Camila Penna: mean, Andrew, where do I even begin with what could go wrong and risks? I mean, anything involving innovation and such a rapid pace of change, there are many things that keep me awake at night. And look, if I'm just gonna talk about a few things, we want clients using our content and using this data, but the
- 12:24
- Camila Penna: usage needs to remain appropriately governed. We want to be able to have some agreement or control around how our proprietary data is getting used. I think a lot about things like machine readable content risking disintermediating the creators, where we stress a point a lot with our clients is
- 12:44
- Camila Penna: that when we deliver this content as data, we really care about the attribution back to the creator. That a machine doesn't just surface a note. It actually surfaces content that can then lead to conversations, that can then lead to strengthening relationships, that can lead
- 13:04
- Camila Penna: to actual interactions as to why that content mattered or why it was created. And we don't want any of that to go away. On the contrary, we're hoping that it enhances that relationship aspect within the clients and our content creators. Another risk that we think about is
- 13:24
- Camila Penna: efficiency versus differentiation. If everyone has similar AI tooling, productivity alone isn't really an edge, is it? The edge increasingly moves upstream to proprietary information, to insight, to judgment, and differentiated people. And one of the greatest risks we have, honestly, is the need to
- 13:44
- Camila Penna: be agile, the need to be flexible, the need to be quick. I mean, for many years we were talking about how we could deliver content via an FTP. And from there, we built APIs. And almost immediately we're getting clients questioning about how can the delivery be via an MCP? And now we do have
- 14:03
- Camila Penna: MCP connectivity. So what's next? What is going to be the new ask? And that is definitely what keeps us on our toes.
- 14:10
- Edwina Lowe: You talked about the sort of most sophisticated end of that client spectrum constantly asking for more. So what is next? Can you give us a sense of what you're working on to continue to innovate in this space?
- 14:24
- Camila Penna: So I would say that what's next is starting. to come at us quite quickly, which is spoken content being transformed into transcripts. So think of transcripts of expert discussions, conferences, webinars, analyst media. There's an enormous amount of institutional knowledge that is created verbally
- 14:44
- Camila Penna: today that historically disappears. So making it more discoverable, actually with shelf life is quite a valuable thing. So when we think about the next steps of content hub, a big part of it that we will be launching soon is the expansion beyond equities of the content that lays there. We're going to be launching macro, commodities, credits,
- 15:04
- Camila Penna: gonna go across asset classes. Then there's an even bigger question: what if email and chat eventually aren't the primary interface at all? A client actually recently told us that there they're close to having all of their PMs and analysts operating outside of email. They're not going to be using email at all. Wow. That is where they're landing, basically, here.
- 15:24
- Camila Penna: And if a machine becomes an interface between the investment professional and the information they need, what else needs to become a feed? And that's what we're constantly thinking about. But most importantly here, Edwina, really is that we want. to be flexible throughout this journey. We want to be able to deliver to the clients
- 15:43
- Camila Penna: wherever this world is taking us. And that's the most important part here is to be able to listen, to be able to be adaptable, and to be able to change with them.
- 15:53
- Edwina Lowe: Absolutely. So you've talked about this journey. Where are we today?
- 15:58
- Camila Penna: I would say that to use a very simple analogy, we're probably at IT. iPhone 1. And I think the important point here isn't predicting what iPhone 15 is gonna look like. It is recognizing that the interface has fundamentally changed. We know content is becoming data. We know AI is changing how people consume it. What none of
- 16:18
- Camila Penna: us knows yet exactly is where that's gonna lead us. The firms that do best will be the ones that keep listening to clients, protect what makes their content. valuable, and adapt quickly as the technology changes.
- 16:30
- Edwina Lowe: Yeah, I love that analogy, and I can't wait to see how this evolves myself. Thank you so much, Camilla, for joining me today and for sharing your insights. It's going to be really fascinating to see where this journey continues.
- 16:42
- Camila Penna: Definitely. And thank you, Edwina, for your insightful questions. We are very excited to continue with this journey.
- 16:48
- Unidentified speaker (UNKNOWN): Thank
- 16:48
- Edwina Lowe: you to our listeners for tuning into this episode of Making Sense. And with that, we'll close.
- 16:55
- Unidentified speaker (speaker_2): Thanks for listening to JP Morgan's Making Sense. If you've enjoyed this conversation, share your feedback by leaving a comment or review wherever you listen to podcasts. And be sure to follow our channel so you don't miss an episode.
- 17:12
- Unidentified speaker (UNKNOWN): The
- 17:12
- Unidentified speaker (speaker_2): podcast views do not necessarily reflect those of JP Morgan Chase Co. or its affiliates, together JP Morgan, and are not from JP Morgan's research department. They do not constitute recommendations or offers to buy or sell securities. Intended for institutional and professional investors, not retail use, it is for informational purposes only.
- 17:32
- Unidentified speaker (speaker_2): Products and services mentioned may not suit all investors or be available in all jurisdictions. JP JPMorgan Morgan may make markets. and trade and discuss securities and asset classes. Visit www.jpmorgan.com forward slash disclosures forward slash sales and trading disclaimer for more disclaimers and regulatory disclosures. External speakers'
- 17:52
- Unidentified speaker (speaker_2): opinions are personal and not JPMorgan's views. Copyright 2026 JPMorgan J.P. Morgan Chase Co.
- 17:58
- Unidentified speaker (UNKNOWN): ⁇
- 17:58
- Unidentified speaker (speaker_2): All rights reserved.
More from Making Sense
Financing the Future of Defense: Admiral Pierre Vandier on Innovation and Investment
Episode discussion: Admiral Vandier on NATO's shift to protracted warfare and the need for private capital in defense adoption
Speakers’ claims: Admiral Pierre Vandier, NATO's Supreme Allied Commander Transformation, argues that the war in Ukraine has shifted conflict from crisis management to protracted warfare, making industrial production capacity decisive. He notes that military R&D has lagged behind civilian innovation for 30 years, creating a cultural barrier to adopting new technologies like commercial drones. Vandier suggests that NATO's primary challenge is now adoption at scale, requiring a shift from innovation boards to adoption boards. He calls for private capital and banks to inject speed and new KPIs into the defense sector, moving investment from legacy systems to future technologies.
Revving the ETF vehicle: Roundhill’s CEO on DRAM, next-gen innovation and more
Episode discussion: Roundhill CEO Dave Mazza discusses ETF innovation, DRAM exposure, and retail investor influence
Speakers’ claims: Dave Mazza, CEO of Roundhill, explains the firm's mission to serve millennial and Gen Z investors through de-aged financial advice. He highlights MAGS as the first ETF providing precise exposure to the Magnificent 7 via cash equities and total return swaps. Mazza also details the DRAM ETF, which offers direct access to memory semiconductor leaders like Micron and Samsung, addressing the limitations of traditional semiconductor or South Korean equity proxies. He distinguishes between structural themes like AI-driven memory demand and short-term catalysts, noting that competitor launches validate these theses. Finally, Mazza asserts that retail investors now account for 20-25% of US daily volume, up from pre-COVID levels, and influence product development through online communities.