A small group of P&C carriers has moved AI out of pilots and into core operations. The financial gap between that group and everyone else is now measurable in revenue growth and share price. If your carrier sits outside the group, the distance is already visible.
Carrier size is part of the picture. Large carriers run more AI pilots and score higher on governance, and each lead compounds the next.
This article explains what the research shows, what it does not show, and where your own gap sits. It covers the three gaps the research points to: measurement, ownership, and data connection. It closes with a four-step check a mid-size carrier can run on one live use case.
What Is the AI Gap Between Tier 1 Carriers and Everyone Else?
The gap is the distance between the 10% of P&C insurers that run AI as a core operating capability and the majority still piloting it.
"Tier 1" here means the largest carriers in the market. No source we reviewed sets a premium threshold for the term, so this article follows the research and uses "large" and "midsize" the way those studies do.
What the Capgemini data shows
Capgemini's World Property & Casualty Insurance Report 2026 calls the top group "intelligence trailblazers." Over three years, these insurers achieved up to 21% higher revenue growth and an approximately 51% greater increase in share price than their peers.
Those figures rest on insurers' self-assessed AI maturity and self-reported financial results from 2021 through 2024, according to the report's own footnote. They show a strong association between AI maturity and results. They do not prove that AI caused the gap.
Where carrier size enters
Datos Insights surveyed 44 P&C insurer CIOs in the second quarter of 2026. Agentic AI pilots reached 74% of large carriers and 24% of midsize carriers, up from 22% and 4% a year earlier.
The same survey scored governance maturity at 3.9 out of 5 for large carriers and 2.7 for midsize carriers. Both measures cover pilots and governance practices, not production results, and the sample is small.
What the data does not show
No study we reviewed measures market share moving from midsize carriers to large ones. What exists is a projection. Jefferies analysts believe the largest insurers will carry the biggest AI implementation budgets and have strong motivation to defend their positions, which raises the prospect of share consolidating among a smaller pool of large insurers over time.
The preconditions for a shift are in the data. The shift itself is not yet measured.
Why Are Most Carriers Stuck in the Pilot Phase?
Capgemini's data partly explains the pilot-phase stall in P&C: results go unmeasured, ownership stays unclear, and technology spending outruns training.
Results go unmeasured
In the Capgemini survey, 42% of insurers track no AI metrics, and 60% remain in exploration or proof-of-concept work. Another 55% report no clear return on their AI initiatives.
Without a metric and a baseline, a pilot cannot be declared a success or a failure. It simply continues.
Ownership stays unclear
The same share of insurers, 55%, say it is unclear who owns AI at their firm. Two-thirds cite a shortage of AI skills.
Capgemini reports that responsibility then falls to individuals or small teams, which makes firmwide impact impossible. A pilot with no owner has no one accountable for scaling it.
Spending outruns adoption
On average, insurers commit 72% of AI investment to technology and infrastructure and 28% to change management, including training. Trailblazers are nearly four times more likely to invest in change management beyond basic training.
The result shows up on the floor. Among employees with access to AI tools, 47% report that their workday is unchanged after 18 months of use.
How Do Legacy Systems and Data Quality Hold Carriers Back?
Legacy systems and fragmented data hold AI back at the technical layer: 81% of insurers cite legacy architecture as a barrier, and 74% cite data quality and access.
Those figures come from Risk & Insurance's coverage of the Capgemini report, and they describe the insurers surveyed, not carriers of one size.
Legacy systems multiply mapping work
Policy, claims, and billing data often sit in separate systems with different field names, codes, and formats. An AI tool that reads one structure meets a different one in the next system.
Every new data source adds mapping and validation work. That work is manual in many carriers, and it sets the pace of every AI use case that depends on the data.
Data readiness lags the ambition
Only 12% of insurers report very high maturity in data readiness, according to Capgemini, even though insurers rely heavily on unstructured data such as notes, documents, and correspondence.
A pilot built on a hand-cleaned sample works in a demonstration. It stalls when the live feed arrives in a different shape.
How Can a Mid-Size Carrier Find Its Own AI Gap?
The Measure-Own-Connect Check tests three things: whether AI results are measured, whether one executive owns them, and whether data reaches the AI from existing systems.
InsOps built this check from the barriers that Capgemini and Risk & Insurance reported. It is a diagnostic, not a benchmark, and it runs on one use case at a time.
Step | Question to answer | Gap it tests | Where the research appears |
|---|---|---|---|
Measure | Does the use case have a written metric and a baseline? | Unmeasured results | Results go unmeasured |
Own | Is one named executive accountable for the result across units? | Unclear ownership | Ownership stays unclear |
Connect | Does data from policy, claims, and billing systems reach the AI without manual re-keying? | Legacy and data quality | Legacy systems multiply mapping work |
Measure
Write one metric and one baseline for each pilot before it expands. Time from first notice of loss to assignment, or hours spent re-keying one document type, are examples a claims or underwriting lead can pull today.
Own
Name one executive and put the responsibility in writing. Capgemini found that trailblazers are nearly twice as likely to embed AI responsibilities directly in job descriptions.
Connect
List every source system that feeds the use case. Count the fields that need manual reconciliation before the AI can read them. That count is the size of your connection gap.
What Should a Mid-Size Carrier Do First?
Run the Measure-Own-Connect Check on one live use case, fix the weakest of the three gaps first, and expand only after a written metric shows movement.
Capgemini found that trailblazers address strategy, technology, and adoption at the same time rather than in sequence. A single-use-case check keeps that discipline at small scale, because all three gaps are scored together even though one gets fixed first.
The steps are:
Pick one use case already in pilot. Choose the one with the clearest owner and the most accessible data.
Score the three gaps. Rate Measure, Own, and Connect as green, yellow, or red, using the questions in the table above.
Close the weakest gap first. A red on Own needs a named executive. A red on Connect needs a source-system map.
Review the metric, then decide. Expand the use case, adjust it, or stop it, based on the written baseline.
How InsOps Helps
InsOps builds an insurance-trained AI that assists with the data connection gap. LiLa, our insurance-trained LLM, runs inside your own environment, so PII and PHI never leave controlled infrastructure. A person reviews and validates every mapping before it is deployed.
Our Integration Gateway connects to Guidewire PolicyCenter, ClaimCenter, and BillingCenter, so data from your source systems reaches your Guidewire workflow without custom engineering for each source. LiLa maps and transforms incoming payloads into Guidewire structures.
InsOps migrates legacy data into Guidewire and keeps it flowing in real time.
If you are evaluating how to close the data connection gap without waiting on a core replacement, contact us to talk through what this could look like for your operation.
Frequently Asked Questions
What is the AI gap between large and midsize insurers?
It is the difference in how far each group has moved AI beyond pilots. Large carriers run more pilots and report stronger governance, and a small cohort of insurers of all sizes treats AI as a core operating capability.
Why does the AI gap matter for market share?
Capgemini ties advanced AI maturity to faster revenue growth, and Jefferies analysts have raised the prospect of share consolidating among large insurers. The share shift itself has not been measured yet, so the gap is an early signal and not a recorded loss.
How can a midsize insurer start closing the gap?
Start with one live use case and run the Measure-Own-Connect Check on it. Fix the weakest gap first, then expand only after a written metric shows movement against a baseline.
What stops insurers from scaling AI beyond pilots?
Three patterns recur across the research. Results are not tracked, no one is accountable for them, and technology budgets run well ahead of training and change management.
What should an insurer measure to know whether AI is working?
Pick a metric tied to the use case and record the baseline before launch. Time to assignment, rework rates, and hours of manual re-keying are common choices because they can be pulled from existing systems.
How does InsOps assist with the data connection gap?
InsOps builds an insurance-trained AI called LiLa that maps and transforms data from source systems into Guidewire structures. It runs inside your environment, and a person reviews and validates every mapping before deployment.
How long does it take to close the gap?
No study we reviewed publishes a standard timeline, so any fixed number would be invented. A phased start on one use case, with a written baseline, gives you your own timeline within the first review cycle.
Is carrier size the cause of the AI gap?
The research cannot separate size from maturity. Capgemini defines its top group by AI maturity, not size, while Datos shows large carriers ahead on pilots and governance. A midsize carrier can still close the gaps it controls.