Good Morning!
A lot of community banks are planning 2027 in the same order: connect the systems, clean up the data, then fund AI. I'd flip that order, and this week's deep dive explains why.
The regulators gave you useful leverage for that conversation this month. The FDIC, OCC, and Fed said they'll look at core provider contract terms that make it hard for banks to add other services, and the FDIC wants bank mergers approved faster. Both belong in your 2027 planning.
TRENDING AI NEWS FOR CB
Community Banks Should Renegotiate Their Core Contracts This Year
Community banks should walk into the 2027 budget season ready to renegotiate with their core providers, and regulators have just made that easier. On September 11, the FDIC, OCC, and Fed issued a joint statement on community banks' engagement with core service providers. They noted that a few large providers hold a significant share of the core market, which limits community banks' negotiating power. When deciding where to focus supervision, they'll consider contract terms that make it difficult for banks to switch providers or bring in supplementary services.
Why it matters for your bank: An AI agent is only as useful as the systems it's allowed to reach. Ask for open integrations and the right to add other providers now, while the regulators are on record, and don't wait for your renewal date to do it.
If Mergers Take Weeks, Your AI Rollout Should Too
If a bank can now buy scale in a matter of weeks, a bank that plans to stay independent should be able to ship an AI workflow in weeks, too. The FDIC board voted on September 17 to propose a rule that would make merger reviews faster and more predictable. Average review time has already dropped from 107 days in 2023 and 2024 to 64 so far in 2026, and the proposal would add a rapid-processing track that could clear very small deals in as few as five days.
Why it matters for your bank: Pick one workflow for 2027 and hold it to a merger-style timeline: weeks, not quarters. That's scale you can build without a merger.
Agentic Payments Will Favor the Banks Customers Already Trust
Trust will decide who wins when AI agents start moving money, and that favors community banks. On September 22, six banks, including Bank of America, Capital One, ING, and NatWest, published principles for how AI agents should shop and pay on a customer's behalf, covering identity, authorization, fraud prevention, and liability. Consumer data in the same coverage shows only 24% of people would let an AI agent shop and pay for them today.
Why it matters for your bank: The biggest banks are writing these rules now. Put agentic payments on your 2027 watch list, and have your fraud and payments teams read these principles before your customers start asking about them.
DEEP DIVE
The Integration Problem Is Your AI Project
Many 2027 technology plans follow the same order. Connect the systems first. Clean up the data. Then, once the house is in order, fund AI. It sounds responsible, and I think it's backward.
The most valuable thing an AI agent does for a community bank is work across systems that were never built to talk to each other. In one workflow, it can read the loan file, retrieve the required data from the core, and update the loan origination system. That is the integration project, and it pays for itself as it goes, because every workflow you automate leaves cleaner data for the next one. Funding integration in 2027 and AI in 2028 gets you the same result a full budget cycle later.
Bank Director's 2026 Technology Survey, out last week, shows how much is riding on this. Sixty-nine percent of bank leaders call integration a top concern, and more than half of banks that failed a tech project blamed it. I read that as a clear signal of where the AI money should go.
It also changes what you ask your vendors. I'd look closely at whether your core and other incumbents are opening up their software or walling it off. Agents work best when they can move across your systems without borders. Having a vendor today doesn't make them the right partner for the next five to ten years.
Then write the line as a growth line. Three in four bank leaders in that survey see technology mainly as an efficiency tool. Efficiency is real, and growth is the bigger prize. A small business owner who gets a loan decision within hours, not days, remembers who said yes. That's how a community bank competes with a fintech that has never met the customer.
I learned what the wrong-order cost is at Glean. We built the company around invoices and adapted too slowly when payables moved to cards. Ramp bet on cards and raised money at a $44 billion valuation this June.
I think there's far more risk in waiting too long to try, test, and learn than in starting small this year.
Three things to do this week:
Pick one workflow that crosses at least two systems, and make it your AI line. Core to loan origination is a good start. Name the owner and one growth metric, such as time to a loan decision.
Ask your core provider, in writing, what it takes to connect an outside agent to your data. Also ask what your contract says about adding other providers.
Add a tuition line. There's a tuition expense to agentic AI: you have to try, test, and spend some to learn. A small amount set aside for pilots you haven't chosen yet means a good idea in June can start in June.
Related: Our community bank AI playbook walks through where to start.
FROM MULTIMODAL
If You Offload Expertise to AI, You Offload Control
Erik Librader and Ian Butler of Vistrada, a business risk and technology consultancy, work inside dozens of institutions after the AI contract is signed, so they see what happens next. They explain why the right setup is usually a chain of small models rather than a single chatbot, and why handing expertise to AI means handing over decision-making and control, too. If you're setting 2027 AI budgets now, the section on budget, sponsorship, and ownership is worth your time.
What This Actually Means
What did regulators say about core providers?
In a September 11 joint statement, the FDIC, OCC, and Fed said a few large providers limit community banks' negotiating power. When deciding where to focus supervision, they'll consider contract terms that make it hard for banks to switch providers or add supplementary services.
How many community banks use AI today?
Most. Bank Director's 2026 Technology Survey found that 72% have implemented generative AI, 30% use agentic AI, and almost a quarter use neither.
Should integration come before AI in a community bank's budget?
Ankur's view is no. Agents can do the integration work as part of real workflows, and every workflow you automate leaves cleaner data for the next one.
Should AI be budgeted as an efficiency or a growth investment?
Ankur argues for growth. Faster decisions for customers, like a same-day answer on a small-business loan, win business in ways cost savings alone don't.
