In financial services, the most important AI hire isn't a data scientist. It's the owner.
11-16-2026
Experimentation is where most of us started. Real AI tools in people's hands, guardrails up, and the room to discover what these technologies could actually do inside compliance walls. And people amazed us. A task that used to eat three days now takes an afternoon. Someone rebuilds a whole workflow around a use case nobody handed them. The invention is real, and it shows up everywhere: banking, insurance, and wealth management.
The way a person uses AI is deeply personal. The prompt, the judgment of when to trust the output, the instinct for what to feed it, all of it lives in one person's head. And brilliance that remains in one person's head doesn't scale.
The gap was never the technology. We had plenty of tools, and some of them are still sitting in the garage with the plastic on. What was missing was the thing around the technology: an owner and a way to carry a proven workflow from one team to the next. Before Adobe, that owner was me, and it was my job to scale it.
New Adobe research backs this up. Adobe and Incisiv surveyed 5,633 executives across 12 industries; 528 of them came from financial services. The 6% pulling ahead with AI got there by organizing around the work, not by buying better tools. They didn't out-tech anyone. They out-organized everyone else.
Most firms haven't gotten there. 60% sit in what the report calls the Frozen Middle: AI is live inside individual functions, a propensity model in acquisition, a churn tool in lifecycle, but nothing connects, so nobody can compare results, defend them, or account for the full impact. The technology works fine. The operating model around it hasn't caught up. Breaking out of it doesn't require a pickaxe or a full reorganization. It requires a handful of key decisions the Leaders have already made.
Name one owner, or watch good ideas stall between teams.
Only 20% of financial services organizations have a formal AI owner with a title and budget authority. In the rest, accountability is part-time, committee-led, or handed out on a project-by-project basis. That sounds harmless until your lifecycle campaign needs behavioral data from a digital banking platform, relationship manager signals from an advisor portal, and personalization rules shaped by suitability requirements, and no one has the authority to say yes across all three. So, the work sits.
Leaders settle this first. They're 4.7 times more likely than the wider industry to have named a single accountable owner. This isn't a reorganization. Reporting lines stay where they are. It's one person, ideally someone whose whole job is managing your AI program. Give that person the budget and the mandate to set priorities, resolve conflicts across risk and compliance, measure impact, and tell the story that unlocks the next round of funding. Ownership is what makes scaling possible. Without it, you'll be in pilot purgatory.
Put every 'no' that matters in the room early.
Out of every 100 AI ideas proposed in financial services, 15 reach production. Security review alone removes 32 at a single gate. Each of those checks is necessary. Model risk, use case review, data protection, regulatory obligations — all of it protects the customer and the firm. The trouble is timing. When requirements surface only during a formal review, teams either rebuild the use case or abandon it, and the outcome they wanted is no longer what they intended.
It's the same story with technology. Three-quarters of financial services firms still run business and IT as a handoff: the business files the request, and IT responds on its own timeline. That worked for slower cycles. It breaks on AI, where data access, integration, and measurement all have to be solved while the work is moving.
Leaders change the order. They bring the people whose "no" can stop the work in the room during ideation, not at the finish line. The ones doing this well have a documented process where responsible AI, technology, risk, and legal collaborate on a set timeline, making "in the room early" a standard practice. Among Leaders, 81% work in collaborative or integrated relationships, compared to 25% of the wider industry. A "no" you hear at the whiteboard costs you a conversation. The same "no" at the finish line costs you the entire build.
Your believers move more people than your mandate.
Depending on your organization, getting a new AI initiative approved can feel like a Ninja Warrior-style obstacle course. When that AI initiative underperforms, the instinct is to protect the deadline and bury it. Organizations should avoid doing this as this is a costly mistake. can lead to future disasters. 29% of financial services organizations show destructive failure patterns such as assigning blame or repeating the same mistake. Among Leaders, the figure is under 1%. They ask what happened, why it happened, and what must change before the next deployment, and they talk about it openly in operating reviews and steering conversations. Treat it like a data point, add it to your lessons-learned file, and move on.
Then they let practitioners lead. Among Leaders, 42% identify finding and backing internal champions as their single most effective cultural action, ahead of executive sponsorship at 16% and formal training at 11%. These are the lifecycle analysts, acquisition managers, and wealth marketing teams that have already made AI work within real controls. Give them recognition, a small budget, and the authority to carry a proven workflow to an adjacent team. Pair their results with playbooks, evaluation methods, and governance guidance so the lessons can transfer. Adoption spreads because people trust a colleague's results more than a mandate from the center. This is the payoff of naming an owner in the first place: someone whose job is to find the wins hidden within individual workflows and give them a way to scale.
Where to start?
None of this needs a task force. Pick the one problem that matters most right now and put these four moves around it.
- Name the owner. One person, a real budget, the authority to say yes across teams. Everything else depends on this.
- Get the "no’s" in the room early. Bring responsible AI, risk, legal, and technology partners into the discussion while the idea is still on the whiteboard, and not after it's built.
- Set the baseline before you launch. Start by identifying the current metrics such as conversion rate, retention, share of wallet, speed to market, and advisor engagement. And if you can, attribute it to revenue. What happened when you sped up the process? How much better did that creative perform compared to a similar campaign last year? Follow the story: you launched, but what impact did that launch actually have? If you need to collect the data manually in spreadsheets, do it. It's hard at the beginning, but the payoff is real. Money talks, and funding is everything in this AI game.
- Back the person who already made it work. Find your champion, fund them, and give them the room to carry the workflow to the next team.
That's how financial services organizations leave the Frozen Middle. Not with a bigger model. With a smaller, sharper set of decisions made in the right order.
Read the full report to see where financial services organizations stand and what Leaders did differently. Explore the report.
Recommended for you
https://business.adobe.com/fragments/resources/cards/thank-you-collections/financial-services