In 2026 a product manager at an MGA can type “add windscreen cover to Motor and re-rate it” and review the configured change before lunch, while a carrier down the road still waits a quarter for the same change in its core system. That gap, not the number of apps or portals, is what digitalization in insurance comes down to now. AI is moving from pilots into claims, underwriting and product configuration, and the insurers that benefit are the ones whose platforms let them act on it.
Key takeaways
- AI use is spreading fast but stays shallow: 46% of US independent agencies use it, and 8% have it in daily workflows.
- The working examples share one design: AI drafts, and a person approves anything that affects a policyholder.
- With premium growth slowing to 1.3% in 2026, the cost and speed of changing products matter more than new channels.
- Customizable platforms and specialty products are where digitalization now pays back fastest.
What digitalization in insurance means in 2026
Digitization turns paper into data: scanned forms, PDFs, online applications. Digitalization changes how the work is done once the data exists: rules that run automatically, claims that pass straight through, products changed by configuration instead of code. Most insurers finished the first long ago. The second is where they still differ.
For an insurer, MGA or broker the practical test is simple: how long does it take to change a product, settle a simple claim or answer a question about the book? The answers depend less on how many digital tools a company has bought than on whether its core systems let those tools act. Our guide to insurance software explains what each of those systems does.
Digitalization in insurance: 2026 statistics
Technology has raised productivity but not yet profits. McKinsey estimates that labour productivity rose 14% in P&C and 24% in life across claims, servicing and policy issuance, while premiums grew about 4.9% a year since 2005 to roughly $8.3 trillion in 2025. The gains were offset by rising IT costs, compliance overhead and the complexity of layering digital tools onto legacy operations (McKinsey, July 2026).
Growth is slowing. Swiss Re Institute expects global premiums to grow 1.3% in real terms in 2026, down from 3.9% in 2025, with non-life at 0.6% (Swiss Re Institute, July 2026). Customers are going digital where it helps them: in J.D. Power's 2026 property claims study, 49% of claimants used digital tools to submit photos and 45% to receive updates (J.D. Power, March 2026).
AI in insurance in 2026: what works and what it takes
AI tops most insurance agendas this year. The evidence shows broad adoption, a few deep deployments, and a consistent pattern in the ones that work.
Adoption is broad, depth is rare
Among US independent agencies, AI use tripled in two years. The 2026 Agency Universe Study by the Big “I” and Future One finds 46% of agencies using AI, up from 15% in 2024, mostly for marketing content, coverage form analysis and contract reviews.
Depth is another matter. The Big “I” Agents Council for Technology finds that only 8% of agencies have AI integrated into daily workflows, and 55% have no written AI policy (2026 Tech Trends Report). Large insurers look similar. The 2026 Evident AI Index for Insurance, which ranks 30 of the largest insurers in North America and Europe, finds that 49% of their disclosed AI use cases are narrow, aimed at speed, cost and process efficiency, and only 8% point towards agentic reasoning and connected workflows. AI-specialist roles at those insurers grew 32% in a year, to almost one in 50 employees, while their overall workforce shrank 2.2%. For brokers choosing their tools, our guide to software for insurance brokers compares agency management systems, CRMs and policy platforms.
Example: Allianz automates simple claims with seven AI agents
Allianz's Project Nemo, launched in November 2025, handles food spoilage claims in Australia after natural catastrophes. Seven task-specific AI agents check cover, assess the claim and look for fraud, then pass a summary to a person who makes the payment decision. From filing to that human review takes less than five minutes, and Allianz reports an 80% reduction in claim processing and settlement time.
The design is worth copying more than the technology: a narrow, high-volume claim type, agents with one job each, and a person at the point where money moves.
AI in core system modernization
The less visible use of AI may matter more. McKinsey finds that in most insurance core migrations, rewriting code or configuring the new platform is a small part of the work; the time goes on understanding rules, converting data and reconciling results. On its own projects it puts the productivity gain from agentic AI at 10% to 90% depending on the step, with human approvals at each stage gate (McKinsey, April 2026).
Configuring insurance products with AI: the Openkoda AI Product Builder
Product configuration is where AI changes the most for business teams. Openkoda, an insurance policy administration system built by Stratoflow's sister company, includes an AI Product Builder: a product manager describes a new product or a change in plain English, and the AI drafts a typed change-set of coverages, rating factors, underwriting rules and workflow steps.
The instruction “add windscreen cover to Motor and re-rate it”, for example, becomes a new coverage with its limit and excess, a new rating factor and an underwriting rule that refers customers with repeated glass claims. The AI chooses only from a fixed catalogue of operations and never writes free-form code. Nothing is applied until a person approves the before-and-after view, every change is audited, and a structural change creates a new product version, so policies in force stay on the version they were bound under.
The same platform reads uploaded documents to pre-fill a case for a person to confirm, answers questions about the book through Reporting AI, and connects to AI assistants through an MCP server under each user's permissions.
Guardrails for AI in insurance
In the EU, the AI Act classes AI used for risk assessment and pricing of individuals in life and health insurance as high-risk, with the high-risk obligations applying from December 2027. For every insurer, the practical rule is the one the working examples already follow: AI may draft, read and flag, but a person signs off on decisions that affect a policyholder, and the system keeps the record.
Other trends shaping insurance digitalization in 2026
Customizable platforms replace rigid cores and long custom builds
For years insurers chose between a packaged core that changed on the vendor's schedule and a custom build that took years to reach a first product. A third option has matured: configurable policy administration platforms where products, rates, rules, workflows and documents are configuration that business teams change, with custom modules added only where the business is genuinely different.
The test for such a platform is what happens to a change: whether it needs a release, whether in-force policies are protected by versioning, and whether data and configuration can be exported. Core Travel Insurance went from contract to a live platform on Openkoda in eight weeks (case study); our core platform guide compares the packaged suites.
Specialty products move from spreadsheets to platforms
Specialty lines such as cyber, aviation, pet, event and program business attract insurers and MGAs looking for growth beyond standard lines, and they are where generic cores fit worst. These products are often written by MGAs under delegated authority, change frequently and have small volumes, which has kept many of them on spreadsheets and email.
Configurable platforms make them viable to run properly: one product model covering any line, carrier-ready bordereaux for delegated authority, and launch times measured in weeks. SkyGuard, a life and aviation specialty insurer, runs its policy administration on Openkoda.
Self-service where customers actually want it
Self-service pays off at specific moments: reporting a loss, sending photos and following a claim without a phone call. J.D. Power finds satisfaction higher among claimants who use digital tools at each of those steps. The work behind it is less about the portal than about the claim moving on its own once the photos arrive.
Your strategy for embracing digital transformation in insurance
Most transformation plans fail in the same places: they buy tools before fixing how products change, and they run as one programme instead of a series of releases. Six choices make the difference.
Measure the time from decision to live product change
Pick one number before any platform decision: how many days pass between approving a rate or coverage change and quoting it. It captures the cost of the core system better than any feature list, and it is the number a new platform should cut.
Inventory the rules before you choose the platform
The expensive part of a core change is understanding the product logic nobody wrote down, not the software. Document the rating, underwriting and endorsement rules of your three largest products first; the inventory shortens every later step and shows which platforms can express them.
Move one product line at a time
Run the new platform beside the old one for a single product line, reconcile the results, and switch over with a tested rollback. A big-bang cut-over puts every product at risk on the same day; a staged move limits each mistake to one line.
Put AI behind approval and audit from day one
Write the AI policy before the first use case, not after: what AI may draft, who approves, and where the record is kept. With 55% of agencies having no written policy, it is also a credible differentiator with carriers and regulators.
Keep your data and configuration portable
Ask every vendor how you export your products, rules and policy history, and in what format. A platform you cannot leave sets your prices at renewal, and portability is cheapest to secure before the contract is signed.
Count the cost of every tool you add
McKinsey's finding that IT and compliance costs ate the productivity gains is a warning against adding point solutions. Prefer fewer systems that share one record over more tools joined by integrations, and retire something each time you add something.
New business models and revenue streams in digital insurance
Slower premium growth changes which models pay. Swiss Re Institute's July 2026 forecast puts non-life growth at 0.6% in real terms, far below its long-term trend:
In that market, growth comes from products and distribution rather than rate: specialty programs written by MGAs, embedded cover sold at the point of purchase, and partners who quote and bind through APIs. Each needs a platform that can launch and change products quickly and report to carriers without manual work.
Investors are backing fewer, larger bets. CB Insights counts $2.4 billion of insurtech funding across 107 deals in the second quarter of 2026, the highest quarterly total since 2022, with $1.7 billion of it in eight mega-rounds raised by six companies:
For established insurers, that concentration argues for partnering with the few well-funded specialists and integrating their capabilities rather than building each one in-house.
Where Stratoflow fits
We are the implementation partner for Openkoda: we configure products, integrate the systems around the platform, migrate policy data and support the result. When a packaged suite or your existing core is the better answer, we say so. If you are planning the next step in your digital transformation, our insurance team starts with a scoping call and a written ballpark.