By Vriti Gothi

Covecta has raised $6.5 million to expand its AI-powered platform, aiming to help banks automate workflows, accelerate lending, and free frontline staff from administrative burdens.

Despite years of digital transformation spending, commercial loan applications can still take as long as six months to process, with loan officers spending more than 150 hours on a single case. Financial institutions remain stuck managing disconnected systems from loan origination tools and CRMs to public registers and core banking platforms — forcing staff to juggle tasks that should be seamless.

Covecta’s answer is an “agentic AI” platform that sits on top of existing banking infrastructure. Instead of requiring banks to rip out legacy technology, it integrates with incumbent systems and deploys specialised AI agents that coordinate workflows across departments. The platform is available via web and desktop apps and can be deployed within weeks, offering a plug-and-play alternative to years-long tech overhauls.

The company’s first major client, Metro Bank, has already reported a 60–80% reduction in task completion times since adopting Covecta. The bank says the technology has boosted efficiency, sharpened risk analysis, and improved decision-making.

Founded by Scott Wilson, Ben Thomas, and Abdul Hummaida, Covecta’s leadership brings a mix of industry and technical expertise. Wilson previously scaled revenue at Mambu and helped expand Finastra in the U.S., Thomas spent over a decade advising banks on digital transformation at McKinsey and Accenture, while Hummaida has led AI engineering teams at AppSense and Orgvue.

Backers of the platform say its potential stretches far beyond business lending. Covecta plans to expand into asset management, wealth management, and other areas of financial services, aiming to become what it calls an “AI operating system” for the industry.

The investment marks growing confidence in AI-driven solutions that promise not just process optimisation but a rethink of how financial professionals spend their time. For banks under pressure to improve customer service and reduce costs, the question is no longer whether AI will change financial services but how quickly platforms like Covecta can scale.

A startup founded by former Google DeepMind scientists to improve code with AI has raised $5 million.

Hiverge, based in Cambridge, UK, has developed an AI technology to automate algorithm design for companies, with the aim of surpassing what human engineers can do alone.

The startup was founded by Alhussein Fawzi and Bernardino Romera-Paredes, who worked on Google’s coding agent, AlphaEvolve, during its early stages, along with Hamza Fawzi, a University of Cambridge professor.

“What really differentiates companies is the algorithms they use under the hood,” Alhussein Fawzi, the CEO of Hiverge, told Business Insider. “And so what we are doing at Hiverge is designing smart algorithms that go beyond existing algorithms — automatically.”

Engineers can submit a request to improve code on the startup’s platform, “The Hive,” which it refers to as an “algorithm factory.” It uses a technique called program synthesis, in which AI generates an algorithm to meet clearly specified functions, rather than being told how to create it.

Alhussein Fawzi gave the example of instructing the technology to shorten AI model training times by refining existing algorithms, but it could also apply to other areas, like improving runtime for supply chain software.

Unlike vibe coding, in which anyone can prompt AI to create software, Alhussein Fawzi says Hiverge is focused on helping engineers improve backend code with AI instead of a user creating front-end applications. Or, as the CEO put it, Hiverge is working with verifiable pieces of code that are “not just based on vibes.”

Hiverge has been active since late 2024 but has only just emerged from stealth. It is still experimenting with its business model through proof of concepts, Alhussein Fawzi said, adding that it would most likely license its platform to customers.

Alhussein Fawzi left Google DeepMind in August 2024 after six and a half years, during which he worked on projects like AlphaEvolve, the company’s Gemini-powered agent that’s largely used internally to improve hardware algorithms.

Flying Fish Ventures led Hiverge’s seed funding round, which also attracted investment from Ahren Innovation Capital and Google’s chief scientist, Jeff Dean, among others.

Alhussein Fazwi told Business Insider that the funding process was “quick” and lasted “two to three months” from the first investor meeting to deal close. The startup said it plans to use the fresh funding to accelerate its go-to-market, support product development, and expand its research capabilities.

Here’s an exclusive look at the pitch deck that Hiverge used to raise its $5 million seed round.