The Future of AI in Banking is Becoming Clearer. Do These Three Things Now to Stay on Course
By Ben Udell, Contributor at The Financial Brand
Simple Subscribe
Subscribe Now!
AI is moving faster than most bankers want to admit, and the gap between what’s possible and what’s actually being deployed inside banks and credit unions is widening every quarter.
The recent FIS and Anthropic announcement is a signal worth paying attention to. The major AI providers are turning their focus toward the unique needs of financial services, which is genuinely good news for an industry that lives and breathes data privacy, risk management, and customer or member trust. Back on May 4, Anthropic (the company behind Claude) and FIS gave banks and credit unions a concrete preview of what comes next: AI agents embedded directly into core banking workflows.
Key takeaway: The flagship use cases are important to the productivity and security of banks and credit unions (FIs). Anthropic and FIS have identified foundational opportunities to more deeply introduce and embed AI into their customers’ workflows. Initial deployments are underway, with broader availability planned for the second half of 2026. This is the beginning, with more core providers and vendors undoubtedly getting ready to accelerate this focus.
What this means: .The announcement is a jumpstart for every FI still unsure of its AI roadmap. But it’s also too early to fully appreciate how this development, with more sure to follow, will reshape banking. Arguably, we haven’t reached the first inning of how AI will transform the industry. Either way, it should push FIs to think more critically about their AI strategy in 2026 and beyond.
A large majority of the industry is still barely tapping the true capabilities of AI, has not deployed AI to allow for the inclusion of personally identifiable information (PII), and has not aligned a person or team around this technology. FIS just upped the game with a foundational focus on digital coworkers and embedded workflow agents, and most FIs are nowhere near ready for that shift. The AI adoption and enablement gap will begin to widen even faster between those FIs that are taking AI seriously within their leadership teams, and those that are waiting or simply not spending enough time on this transformational technology.
This announcement creates real reasons for optimism, and equally real reasons for caution. It also reinforces steps you should be taking today to create momentum on your AI journey.
Three Reasons for Optimism
1. It lowers the AI bench-strength barrier for FIs that don’t have one.
Most FIs don’t have AI engineers, experts, trainers, or someone explicitly leading AI by title or job description. They don’t have AI product teams building solutions across the FI. Their AI knowledge often doesn’t extend much past the basics of Microsoft Copilot.
What they do have is a deep relationship with their core provider, a familiar vendor management process, and a shared trust in regulatory soundness. Cores are already the most trusted partner for FI data, including PII, customer transactions, and core systems of record. That existing trust is a foundation other AI vendors simply cannot replicate.
A partnership like this gives FIs an on-ramp to expanding their AI usage. It doesn’t eliminate risk, but it lowers friction with the most important vendor partner FIs have. Friction through risk management and lack of internal knowledge is what kills most banking AI projects.
Key insight: Most FIs aren’t looking for moonshots. They’re looking for practical, responsible, everyday AI that fits inside the way they already work. FIS is bringing functional AI to the mass market of bankers, which feels like a more comfortable entry point and a real opportunity to accelerate AI in FIs responsibly.
2. It marks the beginning of the digital coworker era in banking.
This announcement signals the acceleration into the usage and acceptance of AI agents. Think of an AI agent as an employee. A digital employee. Every employee wants their own assistant or at least help with their tasks.
This announcement supports the transition from AI assistants you chat with to AI agents you delegate work to. The published roadmap, covering credit decisioning, deposit retention, customer onboarding, and fraud prevention, represents core work activities that can be meaningfully improved with AI.
Reality check: This is not the final version and the workflows will evolve with AI, but now many FIs can start to build momentum. The industry needs to gain exposure to these concepts now, before the leap from “AI helps me write” to “AI helps me work” becomes table stakes and that gap for an FI is a competitive disadvantage. Starting with risk and compliance use cases is a smart way for FIs to gain exposure and confidence with AI, especially among employees naturally dispositioned to be risk averse.
3. Industry-specific workflows are where smaller institutions can punch above their weight.
GenAI applied broadly to “work” is one thing. Banking-specific AI workflows are another, and a much more powerful path into understanding and applying AI.
FIS’s initial use cases are a smart introduction to agentic AI. These topics are expensive, painful, and deeply specific to banking. FIs spend an outsized amount of time, energy, and resources on basic regulatory requirements, stare and compare, and routine tasks. That’s exactly the kind of problem where agentic AI, given the right data and guardrails, can move the needle. And because the work happens inside an FI-controlled environment with the data already in place, it sidesteps some of the PII and data-handling concerns that have slowed broader AI adoption.
Key takeaway: The FIs that benefit most from this shift will not be the largest FIs. They will be the ones that learn fastest how to apply new technology while finding meaningful internal regulatory relief. Lightweight solutions that make a meaningful impact are powerful inside these relationships.
Three Reasons to Be Cautious
For all the reasons this announcement is exciting, there are equally important reasons to approach it with caution and thoughtfulness, along with a reality check.
The biggest challenge isn’t the technology, it’s the readiness of the industry to absorb it.
1. Most FIs are still organizationally unprepared for AI, let alone agents.
Most FIs today don’t have:
- A clear owner for AI inside the organization
- A documented AI strategy
- AI expectations written into job descriptions
- A measurable AI adoption goal at any level
- Meaningful AI literacy across the workforce
You cannot simply “turn AI on” and expect adoption to happen, even if your core provider gives you the tools. Many FIs already have software solutions, tools, and workflows on the shelf from their core provider that are not fully utilized or gathering dust. You can’t expect to implement new technology from your core quickly if the people management needs are not addressed. The industry has learned this lesson time and time again with new technology and software purchases.
The rollout process that educates and trains, while also supporting the foundations of change management, will dictate the success of these opportunities. The ownership exists with FIS, other vendors, and FI leadership. Doing this well early will make future AI rollouts easier. Doing it poorly will create skepticism and disengagement.
2. Trust but verify with your core providers.
Cores are critical to the success of our industry, but they’re not known for being innovative. FIs don’t turn to their cores first for new technology because cores are best-of-breed. They do so because cores are familiar, the relationship is established, the price is right, and the implementation path is easier. Core providers also struggle with truly understanding how FIs operate day-to-day, especially at community FIs. That’s a structural reality of running regulated infrastructure for thousands of FIs at once.
Core providers think they know what you need to do your job, but the nuance of not having that real world experience will make or break the actual success of these programs. Can a core provider on the cutting edge of AI explain, in simple terms, how their solutions will impact a deposit ops team? Can they walk through how to enable these solutions inside Microsoft products? Will they align enablement with IT? Can they support the nuanced way each risk team assesses risk and outcomes?
Key takeaway: The last mile, connecting transformative technology to end users deep inside an FI, is the mile that determines real success and ROI. This last mile is the hardest step in the process for a core provider to deploy.
3. The industry still hasn’t fully adopted the technology it already has.
Embedded AI with your core sounds powerful. But many FIs have not maximized the technology already sitting on every employee’s desktop, including non-AI related Microsoft products Word, Excel, and PowerPoint. And that’s before you account for the rest of the software FIs buy, implement, and underutilize. Add industry anxiety and limited regulatory guidance to the mix, and you have a real drag on adoption that many leaders don’t fully appreciate.
Key takeawayAI will not magically fix a weak adoption culture. If your FI struggles to get employees to use the tools they already pay for, layering agents on top of those workflows will not solve the underlying problem. It’s a people and change-management challenge before it’s a technology challenge.
Three Things FIs Should Do Now
Here’s what every bank and credit union leader should do in response to this announcement. These steps will help prepare your FI for an AI-enabled future with or without the support of your core provider.
1. Expand your AI exposure beyond ChatGPT and Copilot.
Most bankers work with one model, in one product, for one type of task. There’s not an organizational depth to usage that maximizes performance. That’s not AI literacy or a foundation for the future. Try Claude. Try Gemini. Try ChatGPT. Test the same prompt across models. Let users in a department find the solution that works best for their workflows. Accept that being only a Microsoft Copilot shop is limiting. The market is moving too quickly to assume any single provider will dominate every use case, and your team’s usage needs to be calibrated across the field. And given that providers like Anthropic are increasingly forming partnerships across financial services, broader fluency virtually guarantees you’ll need to understand all of the major models.
2. Start closing your AI skills gap immediately.
AI adoption succeeds when basic principles of change management are applied. Leadership needs to clearly identify and support use cases, provide recognition, publicly endorse responsible use, and in some cases impose consequences for using AI incorrectly or not at all.
That means putting AI expectations into job descriptions and setting AI literacy goals for every department, and certainly for leaders. Create space for experimentation and micro opportunities of innovation. Momentum matters more than perfection right now, and the FIs that build organizational muscle this year will compound that advantage every quarter after.
3. Use trusted environments, knowing AI will continue to evolve.
FIs shouldn’t blindly deploy AI agents, but they shouldn’t wait passively either. There’s a real balance between going “all in” on a solution that might be surpassed six to twelve months from now, and doing nothing while getting bogged down in the bureaucracy FIs often have around new vendors, technologies, and costs. Wait too long and you’ll have to start over. But even if you have to start over, you’re still creating a responsible AI foundation, learning, and building momentum.
AI literacy is best learned by doing and exploring. Responsible people working on responsible solutions are invaluable to understanding how AI works, and how it can work for your FI. Build the muscle of governed experimentation before you need it for something material.
The real value at this stage may not be immediate ROI from any specific tool. The real value is building organizational readiness for what comes next, because what comes next is going to come fast.
The Bottom Line
Whether Anthropic and FIS ultimately dominate this space is beside the point. Banking is beginning to move from isolated AI experimentation toward embedded AI workflows and digital coworkers. AI opportunities will be more readily available. In some cases, opportunities will be embedded and your employees will start to use them unknowingly. This transition is coming faster than most FIs realize.
The FIs that begin learning now, responsibly and practically, will be in a meaningfully different position twelve months from now than the FIs still waiting for perfect certainty. The gap between professionals and FIs adopting AI will compound quietly, quarter after quarter, until one day your customers or members, your employees, and your competitors will all be operating on a completely different playing field than you are. Be the professional and organization that’s in a leadershp position.
