AI Won't Turn Novices Into Market Wizards - But It Can Build Better Brokers
Artificial intelligence is rewriting the toolkit brokers offer traders. It isn't turning beginners into outliers. That was the uneasy consensus at the Finance Magnates London Summit 2025, where executives from AI firms, data providers, and a regulated broker debated where growth will actually come from in 2026.
The panel, "Your Broker's Growth Is Elsewhere, 2026 Edition," drilled into a simple question with big consequences: does AI actually improve trading results, or does the real value sit in engagement, retention, and monetisation? The answer, in short: it's the latter-with measurable lift already showing up in usage and revenue metrics.
Moderated by Adam Button of investingLive, the panel included Roy Michaeli (CEO, WNSTN AI), Kieran Duff (Head of UK Growth & Business Development, Darwinex), Braden Dennis (CEO, Fiscal.ai), and Dor Eligula (Co-Founder & CBO, BridgeWise). From left: Adam Button, Roy Michaeli, Kieran Duff, Braden Dennis, and Dor Eligula at FMLS:25.
Traders Lag, Tools Race Ahead
Duff's take was blunt: most retail traders are still stuck on the basics. "I speak to hundreds of traders a week and not one of them really ever mentions AI," he said. "They're still learning how to trade. There's a massive learning curve before you even get to integrating something like AI."
The blocker isn't code-it's psychology. "AI can't help you manage the emotional trauma of losing a lot of money," Duff added. That reality check landed. Markets have always filtered participants down. The open question is whether AI can cushion that funnel, or just make it nicer to be inside it.
Education Versus Emotion
Eligula countered with data. Across 35 million end users at 90 brokers and banks, BridgeWise sees a clear pattern: when users feel informed-even after a loss-they're more likely to return. "When users feel more educated about the decision they made - even if they lost - they are more likely to come back and try again," he said.
In a recent rollout to a 3.5 million-user broker, AI-driven insights tied to live events lifted volume by roughly 15% in both trade count and size. The point isn't just access to data, but interpretation: "Data alone has partial value. You need a layer on top of it."
Personalisation, Not Chatbots
Michaeli pushed back on the fixation with chatbots. The real unlock is contextual support that adapts to the user. "Think of it as a bionic arm," he said. "It's not telling you what to do. It's empowering you with the right data, at the right time, based on your portfolio and preferences."
That approach lets brokers serve different trader archetypes-ratings-driven, technical, long-term fundamentals-without crossing compliance lines. The craft isn't choosing between engagement and compliance. It's building both into the product from the start.
The Data Arms Race
Dennis spotlighted speed. During Nvidia's most recent earnings, Fiscal.ai processed and standardised the full financials in about three minutes-compressing a lag that used to run one to three days. In earnings season, when retail accounts can drop 7-10% in a single move, that speed matters. It changes what you can show users when emotions run hot. For context on Nvidia's financials, see their investor updates here.
He also sees rising demand for specific, operator-level metrics-Spotify premium subscribers, for example-over headline EPS. Long-term investors want the numbers that actually drive the business, not the ones that fit neatly into a template.
Regulation Shapes the Technology
The sharpest disagreement came on model design. Eligula argued that large language models bring opacity that won't fly in regulated finance, so BridgeWise uses smaller, domain-specific models that can be audited and approved. "LLMs are a black box. In our niche, that's not acceptable."
Michaeli's view: compliance is less about model size and more about framing. Don't give advice. Prompt better questions. Keep users within the rules while giving them clarity at the moment of action.
What Growth Really Looks Like
Across the hour, one theme held: growth won't come from turning beginners into outliers. It will come from products that keep users engaged, informed, and emotionally steadier during stress. The commercial impact is already visible-higher retention, more trading activity, and better alignment between broker incentives and user behaviour.
Dennis closed with a reminder that human nature still runs the show. He referenced the familiar observation that the best long-term results often come from accounts that trade less. Patience beats tinkering. For a data-backed view of behaviour's cost, the DALBAR QAIB is a useful reference point here.
The 2026 Broker Playbook: Practical Moves That Pay
- Make AI the guide, not the guru. Deliver context, not calls. Explain the "why," link to source docs, and surface next-best questions instead of recommendations.
- Personalise by archetype. Let users opt into analyst-driven, technical, or fundamental lenses. Keep the UI stable; switch the insight layer.
- Sync to live events. Tie insights to earnings, macro prints, and portfolio-level risk triggers. Engagement follows relevance.
- Measure behaviour, not just P&L. Track churn, reactivation, time-to-first-trade, and portfolio hygiene. Reward good habits in-app.
- Reduce latency everywhere. Earnings parsing in minutes. Standardisation on ingest. Push alerts when the meaning-not just the numbers-changes.
- Stay auditable. Use model architectures and prompts you can explain. Log prompts, outputs, disclaimers, and user actions.
- Mind the line. Keep language educational. Avoid imperatives. Offer ranges, scenarios, and sensitivities rather than a single "answer."
- Highlight unit economics. Show how fees, spreads, and slippage impact outcomes. Teach risk sizing with simple, repeatable frameworks.
- Close the loop. After a trade, show what happened, why, and how it compares to plan. Help users learn from their own decisions.
Bottom Line
AI won't fix fear, greed, or overconfidence. It can, however, give traders better footing and give brokers healthier businesses. If you build for education, timing, and behaviour-within the rules-you'll see the lift where it matters: retention, volume, and lifetime value.
If your team is evaluating AI capabilities for finance, explore curated tools and courses built for this domain here.
Your membership also unlocks:
