Generative artificial intelligence is starting to impact the way payment fintechs create value for their customers. Stripe is now building its sales pitch to Asian tech companies, merchants and corporates around what its services can do if a client plugs in their own language-learning models, such as ChatGPT.
Stripe sells companies payments tools, to enable things such as billing, subscriptions, and checkout. The US-Irish fintech, currently valued at $50 billion, is trying to make headway by pitching payments plus AI as a business transformation.
“CFOs and accountants are seen as a cost-recovery function, providing guardrails to a business,” said Vivek Sharma, lead for revenue and finance automation, in Singapore. “But they can also help generate revenues.”
Stripe hopes to position itself as a ‘revenue growth stack’ in line with CRMs for salespeople or databases for IT departments. This would include a mutually reinforcing set of data sources, from pre-payment to post-payment, a ‘payments flywheel’.
Connecting the dots
The company obviously wants its various payment services to serve as a treasurer’s building blocks, not just to track and execute transactions, but to create insights from interlocking types of data.
“A company can optimize payments and collection,” Sharma said, “but what comes before the payment?” This could be a billing invoice, a subscription service, a shopping website, or a point-of-sale device.
The fintech contends that by building models off this data, the back office’s focus evolves from just efficiency to insights that can impact revenues.
“Finance teams aren’t treating data as a strategic asset,” Sharma said. That’s probably because harnessing data is not easy. Current databases require knowledge of SQL, or structured query language.
The advent of language-learning models, however, make querying a database as easy as typing in a Google search command. LLMs respond to ‘natural’ questions, and respond in kind. Teams don’t need to code in SQL to glean information from a database: they can use an LLM.
Putting this on top of a payments stack – all the various accounting, finance and transaction information – can make it easy to ask what’s going on with a customer or a segment.
“Everyone in the organization can have access to a company’s beating heart of financials, even if you’re not an accountant,” Sharma said. While these tools have been designed for back-office functions, a salesperson or product-development team could use the same information to build a proposition.
The data might just be proprietary, but Stripe wants to pitch itself as a huge database itself, given all the financial data that flows through its pipes.
- Read more:
- Visa seeks Singapore FX license to chase Asia flow
- “StripeGPT” | Paul Harapin, Stripe | DigFin VOX Ep. 59
- Payoneer building payments biz for the meme economy
Are customers in Asia buying Sharma’s pitch? It’s a region where people, including CFOs, are comfortable with mobile technology. That may not be enough.
“We’re early in Asia,” he said, “because of a long legacy of homegrown tools.”
He also acknowledges companies are wary of the pain of a business transformation, particularly when the economy is struggling. There may not be a budget for innovation. LLMs are unreliable and often make things up.
Like many fintechs, Stripe offers its services on a modular basis, such as payments for a subscription, and it tries to cross-sell. As for hallucinating AI, Sharma says this is why the technology won’t replace humans, but it can help them make decisions closer to real-time. “That’s the disruption,” he said.