The AI Gap No One Budgeted For
By David Squibb, President and CEO, Covenir
Most insurance AI rollouts come with a slide for the model, a slide for the integration and a slide for the savings. Very few have a slide for the person who’ll be using it on a Tuesday afternoon when an upset policyholder calls.
That gap showed up clearly in Covenir’s 2026 Insurance Operations Leaders Trends Report. We surveyed 152 U.S. insurance operations decision-makers at carriers, MGAs and insurtechs in February and March of 2026. The short version: AI adoption is speeding up, and the investment in the people running it isn’t keeping pace.
What the numbers say
- 70% of organizations now have AI running in live operations, up from 58% a year ago.
- 20% are cutting training budgets at the same time. Only 7% are actively protecting them.
- 65% of ops leaders say their teams are more stretched than they’ve ever been. Among executives, it’s 91%.
- 88% are still optimistic about the industry’s future, up from 83% in 2025.
Optimistic and exhausted is a very insurance combination. Leaders believe in where things are headed. They’re also asking smaller, busier teams to absorb new tools while the day job keeps ringing.
A training cut is an AI cut
AI in insurance operations rarely works alone. A voice agent hands a caller to a person. A document tool flags an exception for someone to review. A claims model suggests a next step that an adjuster has to accept or override.
Every one of those moments depends on a person who knows when to trust the output, when to question it and how to explain it to a policyholder. If that person hasn’t been trained, the handoff is where things break.
So cutting training while expanding AI doesn’t really save money. It moves the cost somewhere harder to see: rework, escalations, complaints and the slow drip of policyholders who quietly don’t renew.
More budget hasn’t meant less strain
You might expect the organizations with the most money to feel the least pressure. The data says otherwise. Mid-sized carriers with 500 to 999 employees reported the highest budget growth of any segment, and the highest team strain too, at 83%.
That raises a fair question about where the new money is going. If it’s going into tools and not into the people expected to use them, the strain makes sense.
There’s a second signal worth watching. Among the most advanced AI adopters, 54% say headcount is where they plan to cut investment most in 2026. That’s more than five times the rate of less mature peers (11%).
AI isn’t eliminating insurance operations jobs at scale yet. But as that moment gets closer, the people who stay will be the ones handling everything the AI can’t. That’s an argument for training them more, not less.
Three questions for your 2027 plan
- Who owns the handoff? For every AI workflow, name the human step, the person responsible for it and how you’ll measure whether it’s working.
- Is training inside the AI budget? Treat enablement as part of the cost of deployment, not a separate line that gets trimmed first when budgets tighten.
- Can you act on what you’re collecting? 81% of leaders say operational insights are critical, yet 47% either aren’t using their data or can’t turn their dashboards into decisions. AI will only widen that gap if no one’s trained to read the output.
Where outsourcing partners fit
The report also found that 29% of insurers now expect their outsourcing partners to have AI built into their services. Outsourcing is becoming part of the AI strategy, not just a cost line.
If you’re evaluating partners with that in mind, ask more than “what AI do you use?” Ask who trains the people working alongside it, how handoffs are measured and who carries the investment as the technology changes. At Covenir, we manage and invest in the AI behind our clients’ operations and train our teams on those handoffs, so insurers get the upside without carrying the build.
The insurers who pull ahead in 2027 won’t be the ones with the most AI. They’ll be the ones whose people know exactly what to do with it.
Read more: Download the full 2026 Insurance Operations Leaders Trends Report for segment-level breakdowns. For a practical look at AI-human workflows in FNOL and claims, get the free ebook, The AI On-Ramp.
About the author: David Squibb is President and CEO of Covenir, an onshore business process outsourcing provider built for insurance. Covenir supports carriers, MGAs and insurtechs with call center, FNOL and claims, print and mail, and back-office services, and is part of the Constellation Software family.
