The hole between AI ambition and actual enterprise outcomes got here below scrutiny at FintechOS Elevate 2026, as banking and insurance coverage leaders mentioned what it takes to show funding in synthetic intelligence into significant operational enhancements.
Held at Kings Place in London on 14 October, the second annual Elevate convention attracted round 300 registrations and featured greater than 25 audio system, with discussions spanning product innovation, buyer expertise, governance and the challenges of modernising monetary providers.
At last year’s Elevate, audio system warned towards ‘AI theatre’ and pursuing tasks with out clear enterprise aims. Twelve months on, consideration had turned as to whether all that funding in AI was really delivering outcomes.
In her keynote, The AI Reckoning, writer and board advisor Leda Glyptis challenged monetary establishments to look past particular person productiveness positive aspects and the rising variety of AI pilots.
“Focus is a feature, not a bug,” she stated, warning that poorly coordinated initiatives may introduce extra prices and operational dangers.
Her message was that companies have to measure enhancements to their processes and outcomes, whereas addressing current issues with information, abilities and decision-making.
“If you want to be successful at your AI implementations, governance is your only friend,” she informed the viewers.
Bringing AI into on a regular basis operations
FintechOS CEO and co-founder Teo Blidarus opened the convention by highlighting how fragmented merchandise, workflows and information have been stopping monetary establishments from making full use of AI.
“Data fragmentation is typically a consequence of product fragmentation,” he stated, explaining the want to attach merchandise, information and workflows so AI can be utilized throughout on a regular basis banking and insurance coverage operations.
Blidarus additionally stated FintechOS was introducing a brand new method to implementation, after acknowledging that some clients had been “stuck in very complicated upgrades”.
A yr after unveiling FintechOS 8 at Elevate 2025, the firm launched its newest launch, FintechOS 8 Blue Sky, with new AI capabilities designed to assist product growth and on a regular basis banking and insurance coverage operations.
Blidarus and Byron Levin, head of presales and ahead deployed consulting at FintechOS, demonstrated two AI copilots, Dex and Abi, constructed on the similar agentic framework. Dex helps groups configure monetary merchandise, pricing, guidelines and workflows, whereas Abi helps workers and clients with day-to-day operations.
AI can assist, however who makes the choices?
In one demo, Abi guided a buyer by a conversational mortgage utility. Another used a fictional constructing society to indicate how AI may assist develop and worth mortgage merchandise, assess dealer purposes and put together industrial lending circumstances for evaluate.
But the dialogue additionally touched on how a lot decision-making needs to be left to AI.
“The lender, the bank, the financial institution has to own the decision,” stated Joshua Weinstein, co-CEO of Gable Group, throughout a panel dialogue on retail mortgage origination.
He warned that lenders nonetheless want to grasp and clarify how choices are reached, even when AI has completed a lot of the groundwork.
In the insurance coverage classes, FintechOS additionally demonstrated how AI may assist with motor insurance coverage pricing, claims, coverage modifications and cost reconciliation, with workers retaining duty for approving essential choices.
What does real transformation seem like?
In the afternoon, Dharmesh Mistry, CEO of Vision 20/20 Consulting, challenged delegates to consider how AI would possibly basically change their companies, fairly than merely enhance current merchandise and processes.
“It’s your own good numbers you have to be worried about,” he stated. “Nokia had fantastic numbers. Kodak had fantastic numbers. Right until the quarter when they didn’t have good numbers.”
He reminded delegates that banks and insurers had tailored to the arrival of the web and cell expertise, however steered that a number of applied sciences advancing directly offered a distinct problem.
“When technologies come alone, we adapt. When convergence comes, nobody’s got a framework for it because we didn’t imagine these several things coming together.”
Paroma Chatterjee, CEO of Revolut India, mentioned how AI was turning into a part of product growth fairly than one thing thought of individually.
“We don’t look at AI as a different arm of the business,” she stated, explaining how the firm thought of expertise alongside licensing, native market necessities and buyer wants when growing merchandise.
But bringing AI into extra elements of a enterprise doesn’t essentially imply higher outcomes. Rania Svoronou, world head of AI design at Citi, warned about the threat of build up ‘AI debt’ by a number of tasks that enhance particular person buyer interactions with out making a lot distinction to the general expertise.
AI stage 0 of 5: researched, written and edited by Claire Woffenden with out generative AI; solely on a regular basis instruments reminiscent of spelling and grammar checkers have been used. What the levels mean
