The pitch sounds almost too simple: What if everyday investors had the same kind of tireless, algorithmic oversight that hedge funds take for granted?
That's the wager Boston-based Ekpa is making. The fintech startup, fresh out of Y Combinator, has built a platform that deploys AI agents to monitor retail brokerage accounts—parsing news, combing through SEC filings, flagging market shifts—all while the user sleeps, works, or otherwise lives their life. It's not a trading bot. It won't buy or sell anything. Think of it more as a research assistant that never clocks out.
Co-founders Yuga Patel and Anant Asthana, both MIT graduates, framed the problem in straightforward terms when they launched the service: most retail traders simply can't spend eight hours a day glued to earnings calls and regulatory documents. The information asymmetry between Wall Street and Main Street isn't just about capital anymore. It's about attention.
The Platform: Watching, Not Touching
Ekpa's setup hinges on a careful limitation. Users link their brokerage accounts—Robinhood, Fidelity, Schwab, E*TRADE, and two dozen others worldwide—through a third-party connector called SnapTrade. The connection is read-only. Ekpa can see your holdings, but it can't move a dollar or execute a single trade.
For a fintech startup asking users to hand over their portfolio data, that's not just a technical detail. It's the entire trust proposition.
Once connected, the platform's AI agents get to work. They track your positions across U.S. and international brokerages, scanning for relevant market developments. The dashboard surfaces updates, suggests moves, and pitches new stock picks filtered through whatever risk tolerance you've dialed in. Users whose brokers aren't supported can upload holdings manually via CSV—a decidedly low-tech workaround in an otherwise high-tech operation.
The system pulls from recent news, analyst ratings, SEC filings, and economic indicators, though Ekpa hasn't disclosed which data vendors actually power the backend. In financial markets, the quality of your feed often matters as much as the algorithm processing it.
Strategy Lab: Backtesting for the Masses
The free version offers basic portfolio monitoring and alerts. But for $15 a month, Ekpa Plus unlocks what the company calls Strategy Lab—a backtesting tool designed for users who want to experiment with simple trading rules.
The interface isn't sophisticated by institutional standards. Users can define straightforward conditions: buy when a stock drops 3%, take profit at 1%, cut losses at 1%. Then they test those rules against historical data. It's the kind of capability that professional quants have long enjoyed, albeit with far more computational horsepower and granular control.
Ekpa has shared some early traction metrics—$7 million in connected assets, over 6,500 analyses run—self-reported figures that appear as live counters on the homepage without independent audit. The company also claims its system "outperformed the S&P 500 over short time periods (over +1.5% returns in one week)," a statement that raises more questions than it answers. What period? Which strategy? How much risk?
The legal fine print is unambiguous: Ekpa is not a registered investment adviser. It's not a broker-dealer. The analysis, the company says, is research and education. Not advice. Whether that distinction will hold up as AI agents become more sophisticated—and potentially more persuasive—is a question regulators are just beginning to grapple with.
Where This Fits in the Agentic Finance Wave
Ekpa isn't alone in betting that autonomous AI agents will reshape how people engage with markets. But the approaches vary wildly depending on who's being served.
On the institutional side, platforms like Tradeweb's TARA and LTX's BondGPT are building specialized tools for professional trading desks, focusing on credit markets and other asset classes where speed and precision matter enormously. In crypto, where regulation is lighter and risk tolerance higher, startups like Fere AI (which raised $1.3 million earlier this year) let agents actually execute trades on decentralized exchanges.
Ekpa occupies a middle lane. The agents don't pull the trigger, but they do more than passively summarize headlines. They're meant to surface opportunities, issue alerts when something moves in your portfolio, and ideally save you from drowning in the daily deluge of market noise.
Another Y Combinator company, KelAI, is pursuing a similar agent-based model but aiming at institutional funds rather than retail traders. That both companies emerged from the same accelerator batch suggests YC sees something in the agentic research thesis—whether the market agrees remains to be seen.
The Usual Disclaimers (And Why They Matter Here)
Ekpa's terms of service, dated to late July, include the standard warnings: market data "may be delayed or inaccurate," past performance doesn't predict future results, consult a professional before making investment decisions.
Boilerplate language, sure. But it takes on a sharper edge when paired with marketing claims about beating the S&P. If retail users are connecting their portfolios and following AI-generated suggestions—even if the final click is theirs—the line between research and advice can blur fast. Especially when money's involved.
Patel and Asthana were preparing for Y Combinator's Demo Day this fall, where they'll pitch to a room full of venture capitalists who've seen plenty of fintech startups promise to democratize Wall Street. The two-person team is live and charging subscriptions. User funds stay put at their existing brokerages. Ekpa just watches, flags, suggests.
Whether enough retail traders will pay $15 a month for AI-generated portfolio intelligence is an open question. Robinhood and other platforms already offer zero-commission trades, free research, and flashy dashboards. What Ekpa is selling is something subtler: continuous, personalized attention to your specific holdings. The kind of thing that used to require either a financial advisor or an unhealthy amount of screen time.
As autonomous agents become standard equipment in professional finance—and they are, quickly—Ekpa is testing a straightforward hypothesis. Maybe the same technology that helps hedge funds process millions of data points per second can also help an individual investor figure out what to do about that tech stock they bought six months ago.
Or at the very least, alert them when something actually worth their attention happens.
