Lead-Lag Live

The AI Brokerage Era: How Automation, Custom Indexing, and Agentic Tools Are Changing Investing

• Michael A. Gayed, CFA

Share episode

In this episode of Lead-Lag Live, I sit down with Stephen Sikes, Chief Operating Officer at Public, to explore how artificial intelligence is transforming retail investing, brokerage platforms, and portfolio construction.

From AI-powered research to custom-built indexes and the rise of agentic brokerage, Sikes explains how new tools are giving individual investors access to capabilities once reserved for institutions and what risks and responsibilities come with that shift.

In this episode:
– How AI research tools are changing how retail investors analyze stocks
– What it means to build custom indexes using natural language
– The rise of agentic brokerage and automated portfolio management
– Where AI can help investors and where human judgment still matters
– Why the next generation of brokerages may look nothing like the last

Lead-Lag Live brings you inside conversations with the financial thinkers who shape markets. Subscribe for interviews that go deeper than the noise.

Start your adventure with TableTalk Friday: A D&D Podcast at the link below or wherever you get your podcasts!
Youtube: https://youtube.com/playlist?list=PLgB6B-mAeWlPM9KzGJ2O4cU0-m5lO0lkr&si=W_-jLsiREjyAIgEs
Spotify: https://open.spotify.com/show/75YJ921WGQqUtwxRT71UQB?si=4R6kaAYOTtO2V

Support the show

SPEAKER_01:

I think what Agentic Brokerage lets us do is sort of take all of the power that's underneath the hood, all of the data and capabilities that public really has as a brokerage, and allow our users to, you know, using an AI to take advantage of all of that, whether or not we've decided to sort of build that specific feature. Amazing about the AI advancements that we've seen the last few years is this technology does meet the needs of a really broad set of users, investors across sort of every vertical that it touches.

SPEAKER_00:

That has all the inside jokey slang that you only get if you actually get it. So what's inside? Well, a men's What Up Bleaches hoodie, an exquisite hoodie for her, and a few other things to take you from, I think I get it, to Few Crew certified. Now, if you want in, here's the deal. You have to follow at leadlegrecord on X. Follow me, Mela underscore Schaefer on X. Subscribe to Lead Leg Media on YouTube, and like and share this video. You do that, and boom, you're entered. No gimmicks, no funnels, and no nonsense. One winner gets the whole package. The rest of you stay fed until next year. Happy holidays from the Few Crew. I'm your host, Melanie Shaper. Welcome to Lead Leg Live. The retail brokerage world is being shaken up again. Just recently, firms are doubling down on AI and not just for research, but to reshape investing entirely. New tools are launching that let everyday investors build custom indexes and manage portfolios with automation. It's a turning point where technology could finally level the playing field between institutions and individuals. My guest today is Stephen Sykes, Chief Operating Officer at public.com. And he's one of the people responsible for bringing AI-powered investing tools directly to retail. Steven, thank you so much for being here today.

SPEAKER_01:

Thanks for having me. Very excited to talk about the topic and introduce public to maybe some of the audience that doesn't know us as well. And uh very, very exciting things on the horizon here.

SPEAKER_00:

That's awesome. So I to get right into it, can you start by walking us through how public.com's AI research platform works and specifically how an investor without a team of analysts uh could use it to research stocks and opportunities that traditionally only big institutions uh could access.

SPEAKER_01:

Yeah, I'd be happy to walk through it. I think the first thing to start from a little bit, maybe some helpful context just on who we serve, who our customers are, and ultimately like what they're looking for from the platform. So, public in general, we're a broad-based investment platform focused on what I would say are serious retail investors, people that want to, you know, be successful long-term traders, successful trade, you know, success, successful long-term investors, successful traders, people that really are focused on doing a good job. And again, not necessarily, you know, the most casual approach. Uh and I think overall, when we when we talk to our customers about what they want from AI, and it over the last three years, we've been experimenting a lot. You know, I think it's three separate things, right? It starts with a research, right? And for us, that's the first question. I think you're jumping off point is like, how do retail investors want to conduct their research using AI? And I think it's a combination of both taking, you know, sort of the broad-based generalizable knowledge that we get from sort of ChatGPT and Claude and combining that with the high-quality market data, fundamental data, alternative data, content and analysis that a sort of brokerage platform or investment platform like Public has. And so in our product Alpha, we've combined both of those things. Really, it's the best of breed sort of state-of-the-art LLM, enriched with all of the real-time markets and capital market content uh that we sort of offer to our investors as a part of the platform. I think being able to sort of converse and talk about and analyze all of that data using AI is a really, is a really fun and interesting approach to enriching somebody's research process. Yeah.

SPEAKER_00:

And as you so public.com has recently rolled out uh something called generated assets, letting users create custom AI indexes. How exactly does that work? And when someone picks a theme or say a thesis, how does AI turn that into something investable?

SPEAKER_01:

Totally. And I think this is the second big thing that we've, you know, again, our customers really want to use AI for research. The second thing is really constructing these sorts of portfolios. And I think that's the second opportunity and the second thing that we built sort of in sequence as we've gone down this AI journey working with our members. I think the generated assets product is awesome. What it does is it allows you to take the kernel of an idea, a thesis, right, and build a well-diversified portfolio across that idea. And it could be anything from, you know, super sort of, I want to say like soft and qualitative, like, hey, I want to only invest in companies that are founder-led, right? Or that, you know, follow this set of principles. And again, you can describe those in natural language. And I'll tell you in a second how it does it, that the system will go and find all of the relevant opportunities that match sort of those principles. You can do that. You can go all the way to the other end of the spectrum, which is like hard financial data and you know, build a portfolio based on really clear screens around how they've been performing, growing, price action, and all of those things. And again, I think it's the combination of both like the LLM, really great at doing, you know, sort of these searches on, you know, more natural language stuff and then the hard financial data, bringing both those to bear to build a well-diversified portfolio. And the real mechanics, I think, are interesting, right? And I think a lot of this has been unlocked as a capability in the last, in the last few years and sort of the advances that we've seen in sort of LLM and AI technology. What really happens is behind the scenes, you put in your, you know, your prompts, your thesis for what you want to invest in. And, you know, the first thing that happens is we decide sort of what's the universe of applicable, you know, stocks that that thesis could apply to. Is it 20 names? Is it 100 names? 300, 500, and then it makes that decision. And then once it sort of knows the universe that it's operating in, it goes and deploys an agent across each of those individual securities to determine whether or not it matches sort of the principles or the thesis that you're investing. And then it pulls all of those that are sort of uh positive. Yes, we do think this aligns with the thesis. Uh, it pulls all of those into the portfolio, and then it decides on the weighting. Again, if you've decided and sort of prompted it to say, hey, I want this market cap weighted, I want this equal weighted, I want this reverse market cap weighted, it'll follow those instructions. Or, you know, it'll sort of decide based on sort of the relevance, uh how it's going to weight sort of those individual positions across the overall portfolio. And then once you have that portfolio, you've effectively created your own custom index. From there, it's just, you know, like a direct indexing flow. You click a button to buy that sort of portfolio and it gets managed as a whole unit over time by public.

SPEAKER_00:

So with that reflexibility, though, there must be an increased amount of risk. What are some of the misunderstandings that you see amongst users when they're building these uh custom, custom portfolios?

SPEAKER_01:

Yeah, I mean, listen, I think it's the the risks all come back into the same idea of like, you know, using AI generally. It's not perfect. It's not a perfect technology. It doesn't always perfectly interpret your intent. It doesn't always gather the perfect information. And so I think, you know, again, there's an important step in the process, which is like that iteration, the human review, both continuing to converse to refine exactly what you want out of your generated asset. And then once you sort of have it and you've gotten into a place that you really like it uh with the system itself, the next screen sort of lets you make a more granular change. Like, oh man, okay, I kept prompting the system and it kept giving me Apple. I really didn't want Apple in the portfolio. The system really thought it should be there. The next screen, you click the button and you knock Apple out of your portfolio before you go and implement, right? Or, hey, actually, I think Apple's a little too heavily weighted here. I want to knock that down. So again, I think it's all like again, I think AI in 2025 uh is really about that sort of collaboration and iteration and using it as a tool and recognizing its strengths and weaknesses. And I think that's, you know, sort of a principle under which we've designed the generated assets platform.

SPEAKER_00:

So, uh Steven, can you talk a little bit about Public's laid-out vision of agentic brokerage where research and creation and eventually automated portfolio manager are handled all by AI?

SPEAKER_01:

Yeah, I think again, I think it's sort of following on this sort of pathway from initial research to portfolio construction to really sort of operating the entire brokerage and really executing transactions uh using sort of agentic technology. Um I think really our aspiration is that our our members, our users can use AI to execute complex transactions that would be really onerous or almost impossible to do just using sort of the base layer UI. So I think like stepping back, like an interesting way to think about agentic brokers, and the way we're thinking about it, is you know, to design, you know, with the current public platform that we have, every action one of our members can take is something that we've thought of before. Somebody in our software development team has coded it, has built the button that lets the person do the thing that they want to do, right? They build the drop-down menu or build this option or this feature or interpret this data to build this chart in a certain way. All of those things are fabulous. We have an amazing team that's built an awesome platform to meet our customers' needs, but it's one platform, right? I think what AgenTic Brokerage lets us do is sort of take all of the power that's underneath the hood, all of the data and capabilities that public really has as a brokerage and allow our users to, you know, using an AI to take advantage of all of that, whether or not we've decided to sort of build that specific feature. A great example I like to use a lot, and this is one that I'm really looking forward to because I get a lot of questions for our members, is performance reporting about their portfolio. Everyone has a slightly different way that they want to see the performance of their investment account, right? And I totally understand. Some people want time-weighted returns, some people want dollar weighted returns, some people just want to look at individual positions in different slices, and really cutting that data up is quite complex. And so we've we've built it in a way that it, you know, that that our agent, the agentic brokerage, can take instructions from the customer, the member on exactly how they want to see their portfolio performance and give it back to them. And I think those are just one example of an opportunity. Again, I think we can go into sort of transaction flows as well. So, hey, you say you have a very specific, you know, if this then that type rule that you want to implement in terms of making a trade or selling a position, you can go and do exactly that uh, you know, using the agentic brokerage.

SPEAKER_00:

So, Steven, is there a learning curve involved for uh investors who are coming in with different amounts of knowledge and tech?

SPEAKER_01:

I mean, I think when you talk about things like generated assets and agentic brokerage, really all of the AI-enabled features that we're talking about. You know, I think the brilliant thing is that it's all built to be very conversant, conversational, iterative with the tools, right? I don't think it's one of those things where, hey, you really have to, again, I think you have to understand some of the basic principles about what AI is good at, what it's not, understand what capabilities it has, what it doesn't, frankly, like what we've given it access to and what we haven't and how to use that. So that's probably the biggest part of the learning curve. But I think the wonderful thing about the AI technology is it is somewhat generalizable. And that's the idea, is that it can do more or less whatever you ask it to. And that could be from, you know, the most sophisticated, you know, tax analysis and, you know, uh cash raising uh um sort of approach. Say, hey, I wanna, I need$100,000 for my down payment on my house. I'm invested in a bunch of stocks. Please, you know, sell positions to have the lowest tax consequence, right? Like you could tell it to do that. Quite a sophisticated request and one that we like hear from our members, things that they use advisors for all the time, all the way to the other side, which is, hey, I want you to set a very specific uh price alert on a specific stock or hey, two separate stocks that could be in relation. Hey, when when Google goes up by X and Apple goes down by Y, send me an alert, right? There are things like that that like, again, I think cover the full spectrum of investor needs. And I think, you know, like I said earlier, like the brilliance of a genetic brokerage is that it allows us to meet both of those needs uh without having to sort of purpose build every single feature into the public UI.

SPEAKER_00:

So is there a target client uh for this type type of technology? I mean, does it who does it benefit the most? Is it for buy and hold retail investors or traders or or advisors?

SPEAKER_01:

I think that's the one, I think that's the wonderful thing, right? I think it can be an incredible tool for both sort of long-term buy and hold, fundamentally positive investors that just want to use things like generated assets and euthogenic brokerage to have sort of better after tax, after fee outcomes, right? I think they're very powerful tools there all the way to the other side of advanced traders that want to use it to build sort of like OCO or one cancel other order types and some of the more advanced order types that, again, might have to be nested within an order management system or an OMS on other platforms. Agentic brokerage allows us to bring those forward and really cater to sort of the very specific, you know, order parameters that a given, a given trader or given investor might want. Um, and I think that's sort of the brilliance of it. I think that's what's so fast, you know, so amazing about the AI advancements that we've seen the last few years is this technology does meet the needs of a really broad set of users, investors across sort of every vertical that it touches.

SPEAKER_00:

So just lastly, Stephen, for those who want to go and check out these new tools and features, where's the best place for them to go? And what's the easiest way for them to test it out?

SPEAKER_01:

Yeah, listen, I think I always tell people the same thing. Get started on public, open an account, public.com slash sign up, public.com. You could just go browse, see what the see what the platform's all about, see everything that we offer, sign up. It'll take three minutes. And I think we automatically approve something like 95% of customers. So it's a very quick pass in. Then you can sort of see everything, right? You can you can play around with generated assets, you can see some of the sort of materials we've published on a Gentec brokerage. You can sort of play around with uh, you know, all of the platform and flows again on iOS, on Android, on the web, whatever, whatever platform you prefer. I would just encourage people to get started. And then once you've done that, come and find us. I think you know, I make myself really available across our, you know, Twitter, our public subreddit, which is r slash public app. I'm an easy guy to find, and we love to talk to people uh that are either using public or consider using public about the platform and see how we can help people get started.

SPEAKER_00:

It's awesome. Well, Steven, I really appreciate you coming on and laying this all out for us. And thanks to everyone for watching. Be sure to like, follow, and subscribe for more episodes of Lead Like Live.

Head Shot

Michael A. Gayed, CFA

Host