OpenAI’s dominance in the top-tier banking sector is reducing as other large language model (LLM) suppliers increase market share, according to benchmarking company Evident.
The latest figures from Evident show that OpenAI provides the underlying technology in a third of AI use cases across the largest 50 banks in the world, compared to half the amount last year.
Anthropic’s Claude and Google’s Gemini are two LLMs that are gaining ground among the world’s biggest banks, according to Alexandra Mousavizadeh, co-founder and CEO of Evident: “We noticed that there was this slight drop off of the use of OpenAI and we think it’s a natural diversification that’s going on.”
She said banks are asking for more from the supplier community, which is giving other suppliers opportunities to build banking relationships.
“The competition is really being fought on who can provide the best partnership with the banks which are starting to be more demanding,” she said, adding that banks are increasingly asking suppliers to send deploying engineers into their bank to help them build the architecture, for example.
“It’s no secret that many leading banks favour model agnosticism, which could account for some of the shift away from OpenAI,” Mousavizadeh added. “However, many of the bankers we’ve spoken to in recent months have told us that they’re impressed by what Anthropic and Google are delivering.
“OpenAI still boasts successful partnerships with banks like BBVA and Morgan Stanley and retains a strong footprint in banking, so it will be interesting to see how the AI giant evolves its enterprise offering in the coming months.”
On the market, she said there are more use cases being tested by banks and value creation is happening in what she described as “internal focused use cases”, but she added that revenues are not being increased through the technology yet.
“The leading banks are starting to see it have impact with cost reductions and productivity gains,” she said.
Mousavizadeh warned that in the UK, enterprise adoption of AI is not doing as well as the US and is slipping further behind. She said that while HSBC is in the top 10 globally in terms of AI adoption, it is the only UK representative there.
“It’s not that UK banks are not focused, because like all banks in that top 20 cohort, [they] are really doubling down – but at the top end, they are tripling down,” she added.
“The UK has a really good startup community, but these businesses leave and they take their talent with them. So, there is a weakening ecosystem in the UK and that will have significant impact on all the enterprises in the UK and their ability to adopt.”
Separately, Mousavizadeh said she supports the UK government’s recent announcement that it has appointed two senior tech executives from two banks to champion and guide the adoption of AI in the UK banking sector.
“Putting people from technology communities with tech backgrounds into the regulatory space is a good idea because they understand the risks, limitations and opportunities from a technology perspective, and what the regulators can do and how they can respond,” she said.