From Efficiency to Alpha: How AI Is Reshaping Fixed Income

At the Fixed Income Leaders Summit (FILS) Europe, Eugene Grinberg, Co-Founder and CEO of SOLVE, joined industry peers from ICMA, Credit Agricole and H2O Asset Management to discuss how AI is moving from theoretical promise to practical value in fixed income. The conversation highlighted a key inflection point: firms are no longer experimenting with AI in isolation; they’re embedding it directly into the trading workflow to enhance execution, transparency, and ultimately, performance. 

For Grinberg, AI’s role extends beyond efficiency gains. It’s about transforming how traders and portfolio managers make decisions in real time, whether by predicting trade levels in illiquid markets or improving the way firms aggregate and analyze fragmented data. “Step one was munis, step two is corporate bond predictive pricing, which we launched in June, and enhanced in September with greater coverage and improved accuracy.”  

That progression, from using AI to make sense of unstructured data to generating predictive insights, illustrates how machine learning is being applied across both high-touch and low-touch trading strategies. For high-touch, the focus is on speed and confidence: surfacing pricing intelligence from millions of quotes so traders can quickly identify best levels and avoid being picked off. For low-touch, AI models are fueling data-hungry algorithms, helping firms refine pricing and capture incremental alpha. 

At SOLVE, the predictive pricing model is built on a foundation of proprietary data, an edge that matters in an opaque market. As Grinberg noted in the recent TabbForum AI Report, “We’re training our models not just on publicly available data, but on many millions of proprietary quotes we collect by leveraging other AI techniques to aggregate a tremendous amount of quotes data from unstructured messages between market participants.” This approach has allowed SOLVE to achieve high predictive accuracy and deliver confidence scores alongside each price level, bringing greater transparency to model-driven pricing. 

Still, Grinberg noted that not every firm can, or should, build this kind of capability in-house. A lot of firms eventually realize that the vendor ecosystem is much better positioned to provide these types of solutions, especially firms that don’t have large budgets to invest in specialized talent or infrastructure.

The takeaway: collaboration between market participants and AI-native vendors will be key to accelerating adoption. 

The panel also explored the next frontier: how AI could influence market structure itself. As transparency regimes evolve, and as automation reshapes quoting behavior, the distinction between buy- and sell-side is blurring. “There’s a lot more systematic quoting across the buy and sell side,” said Grinberg. “The more advanced firms are already relying on a combination of people, data, automation, and algos.”  

AI’s impact on fixed income is no longer hypothetical, it’s measurable, practical, and accelerating. From predicting where trades will happen to reshaping investment meetings through transcript analysis and bias detection, the industry is beginning to see how AI can expand human capability rather than replace it. As Grinberg emphasized throughout the discussion, the firms that learn to harness AI’s predictive power today will define the competitive landscape of tomorrow. 

Explore more on AI in Fixed Income in this recent TabbForum report The State of AI in the Capital Markets 2025: Beyond the Hype, from ‘Cool to Core’.