Every bank in the region is running, or looking to run an AI transformation right now, with the vast majority planning to increase AI spend in the coming 12 months. The real question, with all this transformation, is who will your customers meet on the other side of it all?
The case for AI is strong, with multiple headline examples showing just how impactful it can be:
- JPMorgan’s c $2 billion annual AI spend paying for itself “pound for-pound,” through c150,000 employees using its in-house LLM Suite every week (Fortune, October 2025)
- Bank of America’s ‘Erica’ is reported to be handling over 3 billion client interactions, and their CTIO says AI has cut their IT service-desk call volumes by more than half (BofANewsroom, 2025).
- HSBC’s Georges Elhedery pointing to customer resolution times reducing more than 0% since January 2025.
Closer to home, GCC institutions have collectively committed to more than $30 billion in AI projects, with Emirates NBD and First Abu Dhabi Bank ranking first and third on the Evident AI Index for Banks – Middle East & Africa.
The benefits aren’t just promises anymore. They are showing up in 56 countless proof points, with institutions setting out ambitious plans and delivering on them. This is only set to accelerate. So, if it is that much of a slam dunk why isn’t everyone finding every dirham, they have to spend on AI deployment?
Customers don’t like it. The National published a quote in August, “In banking, trust is the ultimate currency”. Yet roll backs post launch, infrastructure break downs and customers feeling automation is motivated by the company’s own cost cutting drive all impact brand value and trust. Customers who are effected don’t always come back.
When the benefits for the organisation are clearly this good, but there is still such a gap in customer experience and engagement, the question becomes how can you deliver scale AI transformation whilst still maintaining your customers’ trust and expectations?
First, let’s look at how the colleagues inside the organisations feel about the shift. Mercer’s 2026 Global Talent Trends found 40% of employees now fear losing their job to AI, up from 28% just two years ago. How can you expect customers to trust a change that your own teams don’t?
Gallup’s data is even more resounding. Their research showed that engagement in the teams they met stayed flat at 31% as AI access was rolled out. But when the roll out was paired with clear user cases, supported through a clear plan and managerial support, that engagement climbed to 53%. It’s clear that technology isn’t the only variable and leaders need to view this as a transformation of their human systems as well.
Transformation used to be described, quite deliberately, as changing processes, capabilities, mindsets, behaviours and belief systems all together. It was never just not swapping in a new system and hoping the rest followed. So, has that much really changed in transformation or is it just that our underlying expectations have?
If you’re sat in the C-suite, scoping your next phase of AI transformation, ask yourself who is the human that you want your customers and your colleagues to meet on the other side of the transformation.
At AcuityX we believe in five things that are worth considering before the technology roadmap decides them for you:
Design the human experience first, then decide what the AI delivers -not the other way round.
The fact that AI can do something is not permission to skip the harder question of what your customers or colleagues actually need. Build prototypes around the real opportunities and pain points, road-test the hypothesis before you build the platform, and go and listen, properly to customers and colleagues before you decide anything. That’s where the guardrails for lasting value get set, not just in the tech spec.
"AI transformations are an operating-model shift, not a systems upgrade."
Build a visible, easy route to a person into the design from day one, not as damage control after launch.
Walk the journey you’re changing, end to end, relentlessly. A 90% resolution rate on
a fully automated process sounds like success until you ask the only question that matters: where did the other 10% go, and who’s waiting on the other end of it? Customers expect an opt out and will feel better about the automation if they know they have the option. It’s up to your design to drive the adoption
Redesign roles, training and reward alongside the technology, not after the backlash forces you to.
AI transformations are an operating-model shift, not a systems upgrade. That means how you structure teams, organise work and drive performance has to move at the same pace as the technology itself. Transformations that touch only the tech layer will only ever bank part of the return and maybe none at all.
Tell staff and customers honestly why you’re deploying AI.
If you change one piece of the puzzle you should expect the rest of it to notice. This means you need to make sure that colleagues and customers hear the reason from you, in plain language, before they infer their own. Don’t hide the cost case if that is part of it; cost savings are real and you’re entitled to talk about them. Just pair them with a clear “why” and evidence of the benefit you’re actually delivering to your customer. Silence gets filled with the worst assumptions, and those assumptions are often very expensive to undo.
Measure, measure, measure - engagement, complaints and NPS with the same rigour you apply to cost and productivity.
You need to know what impact your transformation is having on your staff and customers. What we are seeing is that the institutions pulling ahead on AI transformation aren’t the
ones with the flashiest deployment, they are the ones treating the human system as a core component of the transformation, not a by-product of it. Get these five right, and the AI does what it’s actually meant to do: free your people to be more human, not less. That’s the human you want your customers to see.
