Key AI trends in banking for 2026: expert insights from a panel

AI this. AI that. We see it and hear it all over the place.

The noise is deafening. But while the industry buzzes with hype, most banks are still watching from the shore, because their technology simply won't let them move.

At ENGAGE London, Backbase CMO Tim Rutten sat down with three leaders doing the real-life work. They were Marco Martinez (Microsoft), Kanishka Bhattacharya (Bain & Company), and Dave Murphy (Publicis Sapient).

They talked about the hard, messy, real work of deploying AI in banking right now. The code that nobody knows how to read anymore.

The cost of talent versus technology. The shift from helpful chatbots to agents that can do the work.

The panel agreed: The time for "wait and see" is over.

Here are the five things to know about AI in banking for 2026 if you want to scale successfully. These expert insights will help you act with confidence.

1. Legacy code is no longer an excuse

Generative AI has eliminated legacy code as a barrier to banking innovation. Large Language Models can now ingest decades-old, undocumented mainframe code and translate it into human-readable functional specifications.

We all know the story. You run on mainframes. The code was written 30 or 40 years ago.

The people who wrote it retired 10 years ago. The documentation - if it ever existed - is gone.

You're afraid to touch it because you don't know what will break. So you build around it.

You patch it. You add layers of complexity just to avoid opening the box.

Dave Murphy highlighted a massive unlock. Generative AI has changed the economics of modernization completely.

Key capabilities of modern LLMs for legacy systems:

• Code translation: Convert ancient mainframe code into readable specifications
• Dependency mapping: Identify system connections and relationships
• Documentation generation: Create the missing technical documentation

This breakthrough lets you finally understand what your core banking systems actually do.

You can document the undocumented. You can map the dependencies. Once you understand the code, you can dismantle it.

You can move from a "keep the lights on" strategy to a true modernization roadmap.

The shackles of technical debt are heavy. But for the first time in history, we have the key to unlock them and wave goodbye to the legacy constraints.

2. Unstructured data is your biggest untapped asset

Unstructured data - call recordings, PDFs, contracts, and scanned documents - represents banks' biggest untapped asset for AI transformation. These previously unusable data sources now deliver immediate ROI through AI analysis.

You have millions of hours of call center recordings. You have servers full of PDFs, contracts, and scanned documents. In the old world, this was "unstructured data."

It was messy. It was dark.

To analyze it, you needed armies of humans to manually review it file by file.

Martinez pointed out that this is where the immediate ROI lives.

Take compliance as a prime example. Traditional sampling reviews only 3% of call center interactions. Research from PwC on how AI is reshaping banking shows that banks embracing AI could drive up to a 15-percentage-point improvement in their efficiency ratio.

Traditional vs AI compliance checking:

• Traditional: Sample 3% of calls, hope the rest comply
• AI-powered: Analyze 100% of calls, emails, and texts instantly
• Result: Complete compliance visibility in seconds, not weeks

You can feed every single call recording, every text interaction, and every email into the model. You can interrogate the data instantly. You can ask, "Show me every instance where an agent failed to read the disclosure statement," and get an answer in seconds.

The same applies to trade finance. Checking compliance against specific jurisdictions usually takes hours of manual review.

Now? If you know how to prompt, you have it in the blink of an eye.

The data is the same, but your ability to use it has changed completely.

3. Stop building chatbots. Start building agents.

The fundamental difference between chatbots and AI agents is execution capability. Chatbots retrieve information and deflect problems. AI agents execute tasks and solve problems directly.

Most banking chatbots are glorified FAQ pages with a search function. You ask a question, and they give you a generic answer with a link to a policy document.

Murphy warned against falling into the "optimization trap." Using AI to build a "more sophisticated assistant" or a smarter auto-complete might get you a 10% or 15% efficiency bump. But it's missing the point.

The real shift - the one that changes the industry - is the move to Agentic AI.

We're moving from software that retrieves information to software that executes tasks.

Chatbot vs Agent comparison:

• Chatbot: Tells you how to apply for a mortgage
• Agent: Applies for the mortgage for you

Martinez shared a telling story about his five-year-old daughter. She tried to talk to a CD player to change the music. When it didn't answer, she was confused.

She was upset. In her world, if you want technology to do something, you just tell it.

That is no different from your future customer.

They'll expect to tell your bank, "Move $500 to savings and pay my electric bill," and they'll expect the app or agent to just do it.

4. The real cost is talent

Is AI expensive? That depends on how you look at the P&L.

Bhattacharya dropped a necessary reality check on the financials. The cost of the technology itself is plummeting. The cost of tokens and hosting models on Azure or AWS is a fraction of the value they deliver.

According to McKinsey's Global Banking Annual Review 2025, AI could deliver a net 15–20% reduction in banks' aggregate cost base.

Consider the customer service economics:

Cost per interaction:

• Human agent: $24 fully loaded
• AI agent: $0.20-0.40 per resolution
• Savings: 98%+ cost reduction per interaction

But the real cost is the talent.

To build these systems, you need engineers who have built systems at scale. You need "architects and builders" who understand how to integrate new AI models into old banking architectures. Learn more about the AI transformation journey in banking and what it takes to build at scale.

Those people are rare. They're expensive. And you're competing with every other bank (and Big Tech) to hire them.

The bottleneck is your ability to attract the leadership and engineering talent required to execute.

5. Safety comes from experience

There's a lot of fear in the room.

We worry about "deceptive alignment" - where a model pretends to follow your rules. We worry about hallucinations. We worry about the societal impact of automation on jobs.

These are valid concerns. But avoiding the technology just makes you obsolete.

Kanishka compared our current moment to the early days of aviation. Flying became safe because engineers built rigorous checks, redundancies, and protocols. We're on that same curve with AI.

You get safe by building the right guardrails.

For now, that means keeping a "Human in the loop."

AI safety protocols with human oversight:

• Design Time: AI writes code or summaries, humans review before deployment
• Run Time: AI drafts customer responses, humans approve before sending

As confidence grows, the human steps back. But trust is the currency of banking. You can't afford to lose it.

The banks that win will be the ones that move fast with strong controls.

How banks should prepare for AI in 2026

The panel was asked a simple question: "What is the one thing banks should do right now?"

Their answers were direct.

Expert recommendations for immediate action:

• Marco (Microsoft): Start with yourself - treat AI as an "exoskeleton for your brain" and learn to prompt effectively
• Kanishka (Bain): Find a no-brainer use case and execute end-to-end with extreme urgency
• Dave (Publicis Sapient): Reinvent processes entirely with agents rather than optimizing existing workflows

AI waits for no bank. These are the 5 things to know about AI in banking for 2026. The legacy code excuse is dead, and your unstructured data is ready to unlock value.

FAQ: Banking AI Implementation

Q: What's the biggest barrier to AI adoption in banking?
A: Talent acquisition, not technology costs - you need rare engineers who can integrate AI into legacy banking architectures.

Q: How do you ensure AI safety in banking?
A: Start with humans in the loop for both design time (review AI output) and run time (approve AI actions) until confidence builds.

Agents are replacing chatbots. The talent challenge is real, but solvable. And safety comes from moving forward with controls.

The banks that scale AI will have the courage to start and the discipline to execute. They must also be willing to reinvent from the ground up.

So what's stopping you?