Predict, Prepare, Progress: What Insurers Should Expect from ITC Vegas 2026

This blog was coauthored by Theresa Woodiel, Everpure. 

ITC Vegas returns this fall under the theme "The Horizon of Possibilities: Predict, Prepare, Progress." It is a well-chosen framing, because it describes where the industry actually sits. The question is no longer whether artificial intelligence belongs in insurance. It is whether a carrier can put a model inside a regulated workflow; quoting, underwriting, claims, and defend the outcome to a regulator, a reinsurer, a board, and a policyholder.

Organizers expect more than 9,000 attendees, including over 1,900 from carriers. The program is organized around three pillars: Growth, Operational Efficiency, and Strategic Foresight, with curated summits built around problems rather than product categories: AI-Powered Insurance Innovation, The Modern Underwriter, Claims Transformation & Fraud, Cyber Innovation, Commercial & Specialty Lines, Health Connect, and Capacity Connect.

Here's what the 2026 program signals about the industry's priorities and what to press vendors on while you are on the floor.

Agentic AI now has to survive an audit

Agentic systems are moving into submission intake, document triage, quoting, and claims triage, work that sits squarely inside regulated decisioning. The competitive question in 2026 is not whether a model can produce an answer. It is whether you can reproduce that answer, show which data produced it, and show where a human intervened.

Supervisors got there ahead of the market. The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted in December 2023 and taken up by a growing list of states, sets the expectation of a written AI systems program with board-level governance and diligence over third-party models. Colorado goes further for life insurers, requiring a documented governance framework and quantitative testing of external consumer data and information sources. New York DFS Circular Letter No. 7 sets comparable expectations for AI in underwriting and pricing. Carriers operating across states are effectively building to the strictest of them.

The practical implication for ITC: bring your model risk and compliance leads, not only your innovation team. The vendor conversations that matter in 2026 are governance conversations.

Efficiency has become a P&L conversation

Operational Efficiency is its own pillar for a reason. Expense ratios are under pressure, loss-cost inflation has not fully subsided, and agentic workloads are inference-heavy in a way that traditional cloud workloads were not. Cost per decision is now a real line item, and it compounds quietly.

Expect legacy modernization, workflow redesign and overhead reduction to be positioned as innovation in their own right not as the unglamorous prerequisite to it. That is a healthy correction. It also changes the buying question from "what can this model do" to "what does this cost at our volume, and what does it replace."

Claims integrity, not only claims speed

Automated first notice of loss and digital verification continue to compress cycle time. But the same generative tools that speed intake also produce convincing synthetic documents, manipulated damage photography, and cloned voices. Faster intake without provenance checks simply enlarges the fraud surface.

The 2026 claims conversation pairs automation with verification including document authenticity, image forensics, identity assurance. For claims leaders, the metric to watch is not just cycle time and adjuster capacity, but leakage and severity on automated paths versus manual ones.

Growth is a distribution and capacity story

The Growth pillar and the Capacity Connect summit point to where premium is actually being won: modern distribution through wholesalers, agents and brokers; MGA and fronting arrangements matched to capacity; and product design aimed at risks the traditional forms handle poorly such as embedded coverage, parametric triggers, and cyber.

Cyber deserves its own note. The agenda addresses both the weaponization of AI against insureds and the security of insurers' own models and training data. Those are two different exposures with two different owners, and most organizations have only staffed one of them.

Five questions worth asking on the floor

1. Show me the lineage. For a decision this system makes, which data produced it, and can we reproduce it six months from now for an examiner?
2. What is the cost per transaction at our volume? Not model benchmarks but cost per quote, per claim, per submission.
3. Where does a human intervene, and how is that logged? Human oversight that is not recorded is not evidence.
4. How does this map to the NAIC bulletin and our states' requirements? If the answer is a compliance slide rather than a control, keep walking.
5. What does this look like in our environment? Integration with the policy admin and claims systems you actually run, security review, and a realistic timeline.

Where WWT and Everpure fit

Most AI programs in insurance do not stall on the model. They stall on the foundation beneath it; policy, claims, billing, property & casualty and annuity data spread across systems that were never designed to answer a question together, and certainly not in real time.

That is the problem the three of us work on. World Wide Technology brings insurance advisory and integration experience, plus an Advanced Technology Center where carriers can test architectures against their own workloads before they commit budget. Everpure provides the data platform underneath, including performance, efficiency, and predictable economics for AI workloads that are otherwise expensive to run at scale. The platform connects the data layer so the information your models, actuaries, and regulators all depend on is unified, governed, and traceable.

Practically, that means a carrier can answer the lineage question and the cost-per-decision question with the same architecture, rather than bolting governance on after the pilot succeeds.