SBS Embeds AI Across 1,500 Banks, but Full Access Waits Until 2027

SBS has wired generative AI directly into the core banking systems that already run inside more than 1,500 financial institutions worldwide. The vendor, formerly known as Sopra Banking Software, calls the new layer SBS AI Foundation. It targets fraud detection, credit decisions and customer onboarding without banks having to rip out infrastructure they already run.

But broader access will not arrive until early 2027. Separate survey data on the banks that plug into vendors like SBS show most still lack the in-house governance to run AI safely on their own, which is exactly the gap SBS is stepping into.

SBS Wires Generative AI Into Banks’ Existing Core Systems

The system runs on the SBS Data Platform, layering generative AI over data banks already hold instead of pulling information out to a separate tool. Bizcommunity, a South African business and technology news outlet, reported that the design keeps data governed and secured inside each bank’s own environment, so institutions can act on AI-generated insights with the confidence regulators expect.

An AI assistant built into the platform lets staff in service, compliance and operations pull answers straight from internal data. That cuts the admin work and manual searching that slow down everyday decisions, according to the report. SBS says the tool supports both large and mid-sized banks, the same size segment where separate research later in this story finds the thinnest AI governance.

When our clients started asking how to deploy AI safely, and to drive value for their business while meeting regulatory requirements, we built the answer into the platforms on which they already run.

Eric Bierry, chief executive officer at SBS, said in comments reported by Bizcommunity. Bierry added that SBS has spent five decades building the technology and client relationships now feeding the new AI layer.

Why Banks Needed Someone Else to Build It for Them

Banks have not kept pace with fintech rivals on their own. Regulatory complexity, fragmented technology stacks and organizational inertia weigh on incumbents far more than on newer challengers. Roughly 600 AI-focused product launches tracked by McKinsey account for only about 3% of everything the financial sector shipped between late 2022 and the summer of 2025.

Most incumbent banks still use AI mainly to shore up reliability and trust. Treasury automation, chat-based assistants and advisory personalization are useful, but they risk becoming commodity features while fintechs push into faster-growing, revenue-generating categories.

Agentic AI, the term for systems that complete multi-step tasks on their own, like approving a loan or moving funds, rather than just answering a question, is the frontier where fintechs are pulling furthest ahead.

The problem is not that banks refuse to try AI. Nvidia’s 2026 State of AI in Financial Services report found 65% of financial institutions already use it, with nearly 90% either deploying or assessing some form of it. The bigger obstacle now is the sprawl of disconnected systems most banks already built, one tool per problem, none of them talking to each other.

Banks are not standing still either. More than half of banking chief information officers, 55%, said they had already deployed generative AI by the end of 2025, based on a Gartner survey of more than 2,300 banking technology executives. Deploying a tool and running it at scale, though, are different things.

How Many of the 1,500 Banks Are Actually Live?

Not many yet, by SBS’s own account. The company says its AI layer is available now only to a select group of clients, with a broader roll-out across its full institution network planned for early 2027, more than a year after this month’s announcement.

What we know:

• SBS AI Foundation runs on the SBS Data Platform and layers generative AI over core banking, lending and digital banking systems already used by more than 1,500 institutions.
• Access to the new AI layer is currently limited to a select group of clients.
• SBS is targeting a broader roll-out across its full client base for early 2027.

What’s unconfirmed:

• How many of the 1,500-plus institutions have hands-on access to the AI layer today.
• Pricing or licensing terms for banks moving from the core platform to the AI layer.
• Whether the early 2027 target holds across every market SBS serves.

That gap between infrastructure reach and live usage mirrors a pattern across the wider industry. Financial services firms worldwide are moving at a similar pace: 52% were piloting agentic AI or beyond as of early 2026, with less than a quarter reaching the more mature scaling stage.

One Vendor’s Outage Becomes Everyone’s Problem

When a single vendor’s AI layer sits inside the core systems of more than 1,500 lenders, a flaw in that layer does not stay contained to one balance sheet. That kind of concentration is what worries community bank executives most about vendor-embedded AI, according to survey data.

Wolf & Company surveyed 20 executives from 17 community banks holding between $750 million and $25 billion in assets between January 31 and February 28, 2026, the size tier that makes up much of the market SBS and similar vendors serve.

• Security and data leakage – the top challenge for 55% of the community banks surveyed
• Third-party and vendor AI risk – a concern flagged by 40% of respondents
• Limited internal AI expertise – cited by 50% of respondents
• Unclear ROI or business case – a challenge for 30% of respondents

Only 5% of the banks Wolf & Company surveyed had reached a fully scaled, governed AI program of their own, even though 70% said an AI adoption plan was already underway in some form. Just 25% had moved past proofs of concept into actual production.

Banking regulators globally have been explicit about the stakes of leaning on a small number of AI systems. Pablo Hernandez de Cos, chairman of the Basel Committee on Banking Supervision and a former governor of the Bank of Spain, has warned that unchecked models could amplify future banking crises.

Regulators Are Still Catching Up

Regulators are behind the institutions they oversee. Nearly half of financial regulators are still exploring AI or not engaged with it at all, per a Cambridge Centre for Alternative Finance study of 628 organizations across 151 jurisdictions, even though industry adoption runs far ahead.

More than 80% of financial firms said they are adopting AI at some level. The top perceived risk across every group surveyed, industry, vendors and regulators alike, was data privacy and protection, named by 73% of respondents.

U.S. regulators moved on their own front in April 2026, when the Federal Deposit Insurance Corporation, the Federal Reserve and the Office of the Comptroller of the Currency published revised model risk management guidance. That guidance covers traditional financial models but explicitly excludes generative and agentic AI systems from its scope, since regulators say the technology is evolving too quickly to pin down.

The pattern across every row is the same. Adoption headlines move faster than the governance built to back them up.

SBS’s Real Test Arrives in 2027

Two deadlines now sit on the same calendar. High-risk enforcement for AI used in credit scoring under the European Union’s AI Act begins August 2, 2026, adding legal weight to exactly the kind of decision SBS’s platform is built to help automate.

Further out, only one in four banks worldwide are actively using AI to gain a competitive advantage today, according to Boston Consulting Group figures cited in Backbase’s 2026 banking predictions report. The rest remain in fragmented pilots.

Gartner expects multi-agent systems to independently drive 30% of day-to-day banking decisions by the end of 2027, the same year SBS says its full bank network should finally get access to the AI layer it announced this month.