How TresVista Scaled AI Adoption While Cutting Costs with Model ML
5 MIN. READ
About TresVista
TresVista is an enterprise intelligence partner to investment firms and their finance teams. For more than 20 years, it has supported deal teams with research, modelling and analysis, and CFOs with reconciliation and reporting.
Today, TresVista has more than 2,000 professionals and has served over 1,000 clients globally.
The Challenge
As TresVista brought AI deeper into its work, it encountered a challenge shared across the industry: the more useful AI became, the more it cost.
Across consulting and finance, firms have typically responded by capping AI usage per user. This contains costs, but it also limits the value teams can realise from the technology. For TresVista, such limits would have undermined the rationale for adopting AI in the first place.
The Solution
TresVista partnered with Model ML at a moment when both firms were ready to move fast and embrace new ways of working. The teams quickly identified high-impact use cases and built a close, collaborative relationship, solving challenges side by side as the work evolved.
For TresVista, the difference on cost came down to how Model ML approaches the kind of finance work TresVista does. Rather than using the same model for every task, Model ML routes each one to whichever model handles it best, based on accuracy, latency, cost and token consumption. Those decisions are grounded in the Composite, Model ML's benchmark for AI in financial services.
Model ML also runs fine-tuned open-source models on its own GPUs. For the right workloads, this can save up to 95% per customer compared with sending everything to third-party frontier APIs.
With cost under control, TresVista's teams could use as much AI as the work needed, with no usage caps.
Impact
Since January, TresVista's engagement on Model ML is up tenfold and its user base has tripled. Costs didn't follow. They flattened, and in recent months they've come down.
The bigger shift is what that has made possible. With cost no longer a constraint, TresVista’s work with Model ML has grown well beyond its own teams. Agents the two teams built together are now deployed with TresVista's clients, some of the largest financial institutions in the world.
"At TresVista, we pride ourselves on finding the right solutions for clients, and that includes finding the right partners. We believe with Model ML, we found just that," said Sudeep Mishra, Co-CEO and Co-founder of TresVista.
