Executive Summary:
The gap between the best- and worst-run carriers is worth $174 million a year on a $1 billion book—and most of it hides in claims leakage and aging core systems. Your next vendor will promise to close it; the contract you sign decides whether they can. Per-seat pricing ties their revenue to your headcount, so automation that should cut your costs cuts their income instead. This piece gives you three diagnostic questions, three contract models, and a six-dimension framework to find a vendor whose commercial interest is finally pointed at your outcomes not against them.
There is a 17-point gap between the best-run carriers and the worst. Top-quartile insurers run at 17.4%[1] of gross premiums written; the bottom quartile burns through 34.8% and on a $1 billion book, that gap is $174 million a year walking out the door in claims leakage, manual servicing, and processes welded to forty-year-old code. Your next platform vendor will promise to close it. Before you sign, three questions will tell you whether they actually can.
First: does their pricing reward you for closing that gap, or quietly punish you for it? Most core and claims platforms still charge per seat. So, when automation lets you redeploy twenty adjusters, your bill should fall, but their recurring revenue falls with it. You are now paying a vendor whose commercial interest is served by your headcount staying exactly where it is. Every point of expense ratio you fight to recover is a point of revenue they’d rather you didn’t.
Second: can they show you today, before onboarding, where your straight-through processing rate and claims cycle time actually sit against comparable carriers in your lines? Not your own year-over-year trend. Your position against the market. A vendor who can only benchmark you against your own past is managing your perception of progress while your competitors quietly pull ahead on theirs.
Third: when the next cat season hits and the platform underperforms, will they put their own margin at risk or hand you a service credit? A real partner accepts a financial remedy tied to a specific shortfall against a baseline you agreed before signing. A vendor hands you an apology and an invoice.
If even one answer comes back soft, stop. You are not evaluating a strategic partner. You are buying a more expensive version of the exact expense ratio you have today.
How Per-Seat Pricing Costs You Millions
Your last platform RFP probably saved you $400,000 on licensing. It may have cost you $40 million in claims leakage you will not find in the contract. That leakage is not a separate problem from the 17-point expense-ratio gap above; it is one of the largest line items inside it. The $40 million and the $174 million are not two numbers; they are the same fight measured in two ways.
There is a structural conflict built into every per-seat SaaS agreement in insurance technology, and it is almost never named during procurement. Here is how it works: you implement the platform, automate workflows, and reduce operator headcount. Fewer operators means fewer seats. Fewer seats means the vendor’s recurring revenue falls as a direct function of your operational success.
The better the platform performs, the less they earn. That is not a theoretical misalignment. It is a commercial disincentive against full automation, written into the unit economics of the contract you just signed.
Source: https://www.getmonetizely.com/blogs/the-2026-guide-to-saas-ai-and-agentic-pricing-models
Experienced CIOs recognize this dynamic without having named it. It shows up as automation pilots that reach 30%[2] efficiency gains and then stall; features that reduce manual effort but never reach production at scale; quarterly reviews that celebrate adoption metrics rather than measuring seat reduction.
The model that resolves this misalignment is outcome-based pricing: carriers pay for verified business results: claims cycle time, straight-through processing rate, leakage reduction rather than potential system access. When a vendor’s revenue is tied to your claims cycle, they have commercial motivation to push automation as far as it will go. But pricing is only the first lever a carrier can pull, not the whole answer: a vendor can tie its revenue to your outcomes and still control the roadmap, the benchmarks, and the exit. Treat outcome-based pricing as the entry point to a deeper partnership model rather than the destination.
What a Real Partner Looks Like
Only 7%[3] of insurance companies have scaled AI beyond pilot programs, and 70% of scaling failures trace back to people and organizational issues, not technology. The barrier is not finding the right algorithm. It is building the right relationship with a partner who is commercially exposed to your outcome.
Source: https://www.bcg.com/publications/2025/insurance-leads-ai-adoption-now-time-to-scale
Those three questions, pricing, benchmark access, and margin at risk, only get honest answers inside three structural mechanisms. The pillars below are how each question stops being a promise and become: governance you can enforce, reviews with money attached, and AI capability sold on honest timelines.
Pillar 1: Shared Roadmap Governance
A genuine technology partner gives you active influence over their product trajectory, not a seat on a user advisory panel, but formal authority in the process that controls what gets built. The mechanism is a Change Control Board: a joint authority that evaluates every significant change request against cost, timeline, risk, and regulatory exposure before development begins. Organizations that formally integrate project governance with change management are 47%[4] more likely to meet or exceed project objectives than those that manage them separately and adequate executive sponsorship lifts that rate to 71%.
Ask directly:
- Does this vendor provide CCB participation as a contractual deliverable, or only in the sales narrative?
- Who holds authority over product changes that affect your core workflows?
- What is the escalation path when a roadmap priority conflicts with a regulatory freeze window?
Pillar 2: Quarterly Reviews with Financial Consequences
Every vendor pitches the quarterly review as a strategic forum. The distinction is not what the meeting is called; it is whether missing a target costs them money.
A review that means something delivers benchmark data showing where your automation rates sit against the top quartile of comparable carriers, not your own prior-quarter baseline. It proposes specific roadmap adjustments with named owners and timelines. And it carries a direct contractual line between outcome metrics and remediation obligations when targets are missed. Without those three elements, you are managing a support contract at a partnership price.
Pillar 3: AI Capacity with Honest Timelines
Task-specific small language models tuned to insurance operations represent a material productivity opportunity. A hybrid architecture that routes 80–95%[5] of routine queries through localized models, escalating only complex cases to larger frontier models, reduces AI processing costs by over 90% while cutting response latency.
McKinsey deployed 80+[6] AI models for the UK’s largest general insurance firm and reduced liability resolution time by 23 days (from 41 days to 18 days), produced 65% fewer complaints in the affected product lines, and generated approximately £60 million in annual operating savings. That result is instructive, but note the baseline: the gains are only replicable if your starting position and operational complexity are comparable.
“As AI agents take on more of the service workload, customers should expect pricing to reflect the outcomes delivered. We believe customers should pay for the value they realize, not the tools they deploy.”
– Tom Eggemeier, CEO, Zendesk
What a responsible partner will not promise: that this technology will move the majority of your staff from data entry to high-value judgement roles within a single contract cycle. That transformation, where achievable, takes three to five years, significant change management investment, and retraining budgets that belong in your P&L, not a vendor’s product demo. A vendor who presents organizational transformation as a platform feature is describing a decade-long program and handing you the invoice as a sales pitch.
Hold both timelines at once rather than letting them collide: a capable platform establishes a measurable baseline in 30 to 90 days and compounds from there, while moving your workforce from data entry to judgement is the three-to-five-year program. The fast number measures the system; the slow number moves the people.
What Closed Systems and Checkbox Security Really Cost You
40%[7] of employers would switch insurers if their products did not integrate with benefits administration platforms. 90%[8] of insurance executives say future success depends on participating in multi-carrier ecosystems rather than operating in isolation. These are customer retention data points, not technology positions.
McKinsey projects embedded insurance could account for up to 25%[9] of the global insurance market by 2030, making the carriers without API-ready core systems structurally unable to compete for the fastest-growing distribution channel in personal lines. The carriers positioned to capture that share are not the ones with the most product features. They are the ones whose core systems can expose policy, billing, and claims capabilities through secure APIs to external partner environments, without a bespoke integration project for each new relationship.
“Insurance decisions must be explainable, auditable, and defensible. Generative systems introduce outputs that are not always reproducible. That will create a governance problem before it creates a business benefit.”
– Dr. Alistair Prewitt-Crane, Independent Insurance Analyst
Exposing those systems safely, though, usually means moving off the legacy core first, which is the moment the partnership is tested most severely.
Legacy core migration is where most carriers come closest to an operational catastrophe. The ones who navigated it without incident did not uniformly have better technology. They had vendors who ran parallel production environments long enough to validate every data migration assumption, owned the rollback plan as a contractual obligation, and built June-through-November blackout windows for P&C carriers into the project plan by default.
Surviving that migration is a test of operational trust; so is what happens the second something breaks afterward, which is where security stops being paperwork and becomes shared financial risk.
Security as Shared Financial Risk, Not Shared Paperwork
A SOC 2 certification tells you that at the time of the audit, the vendor’s controls met a defined standard. It tells you nothing about their incident response at 2am when your claims portal is down. Financial services breach costs averaged $5.56 million[10] per incident; shadow AI connections alone added $670,000 to average breach costs (IBM).
Active security means continuous threat monitoring embedded in the application stack, automated discovery of shadow IT connections before they become breach vectors, and confirmed vendor vulnerabilities resolved within 48 hours.
Stop Auditing Software Features, Start Auditing Incentives.
The question that separates a real partner from a vendor with a polished security deck: “If your systems contribute to a breach that costs us five million dollars, what is your financial exposure under this agreement?” If the answer is a license credit, the risk is entirely yours.
Bain identifies $100 billion[11] or more in recoverable P&C claims value through AI-enabled workflows, with 30–50% leakage reduction in structured claims deployments. McKinsey estimates automation can reduce the cost of a claims journey by up to 30%,[12] with broader operational transformation delivering 30–50% middle- and back-office cost reduction for carriers that execute the structural changes required. The value is not in dispute. The question is which contract structure gives you the conditions to reach it. The carriers outperforming their peers right now did not find better software. They renegotiated the terms of the relationship. What follows is a structured framework for doing the same.
Three Models, Six Dimensions, One Decision
The comparison below is not designed to be neutral. Per-seat contracts are not neutral either; they are structured to grow with your headcount, not shrink with your efficiency. Use this framework against any vendor conversation. The six dimensions are not features to be ticked; they are the structural terms that will define your cost position, your AI trajectory, and your ability to exit the relationship on your terms.
Out of all the six dimensions, the most underplayed in this context is CLV impact: the dollars a policy generates across its lifetime, not a retention percentage. It is the dimension per-seat contracts quietly cap, because efficiency gains that never reach the customer experience never show up in renewal economics.
| Dimension | Per-Seat SaaS | Outcome-Based Pricing | Continuous Partnership |
|---|---|---|---|
| Time-to-Value | 6–18 months to first measurable workflow change; milestones set by vendor | 3–9 months with delivery tied to defined milestones | 30–90 days to establish baseline; compounding gains from month four onward |
| 5-Year TCO | Rises as integration depth grows; seat count rarely falls despite automation | Scales with value delivered; drops when outcome targets are missed | Highest Year 1 investment; lowest cumulative cost as automation eliminates manual overhead |
| Customization Depth | Vendor roadmap controls product direction; carrier requirements join a queue | Outcome targets create pressure for workflow-specific tuning; limited governance authority | Full CCB participation; roadmap co-owned with contractual veto rights on changes to core workflows |
| AI Readiness | Generic AI features added to standard product; no tuning to carrier’s operational data | AI components tied to outcome metrics; partial tuning possible within contract scope | Hybrid SLM/LLM architecture tuned to the carrier’s claims, underwriting, and servicing data over time |
| Vendor Dependency | Highest; integration debt accumulates with each contract year; migration cost grows | Moderate; outcome baselines create accountability but exit terms remain vendor-controlled | Lowest long-term; workflow logic and model weights contractually owned by the carrier |
| CLV Impact | Marginal; efficiency gains plateau at 30% without roadmap access or governance input | Measurable; 30–50% claims leakage reduction achievable against defined baselines | Highest; compounds as AI models mature against carrier data and governance closes capability gaps continuously |
Segment-Specific Evaluation: The Question Your Contract Must Answer
Everything to this point has been argued through a carrier’s P&L: the expense ratio, the twenty redeployed adjusters, the cat-season remedy. The same six dimensions decide the contract for brokers, MGAs, and the adjuster-facing teams who live inside the platform every day; each segment simply weights them differently.
Each buyer segment in this market carries different exposure across the six dimensions. Carriers weight TCO and AI readiness. MGAs weight vendor dependency and customization depth. Brokers weight CLV impact and time-to-value. Adjuster-facing deployments live or die on AI readiness and claims cycle transparency. The question below is the one question each segment should require a written answer to before signing.
| Segment | Evaluation Question |
|---|---|
| Carriers | Does your vendor’s 5-year TCO model account for the seats you will eliminate as automation scales—or does it assume your headcount stays flat for the life of the contract? |
| Brokers | Can the vendor show you a CLV model—not a retention percentage, but a dollar figure per policy—for what a 20% reduction in servicing time produces on your current book? |
| MGAs | If you terminate this agreement in 36 months, what do you own? The workflow logic, the model weights, the API integrations—or only the exit clause and your data in a proprietary format? |
| Adjusters | What is the baseline claims cycle time this platform was trained against, and what is the documented improvement trajectory for a book of your size, line mix, and jurisdictional complexity? |
The answer that disqualifies a vendor: Any version of “we’ll work with you on that.”
The question above is answerable in a single paragraph by any vendor with genuine operational depth. Ambiguity at the evaluation stage is not a negotiating posture. It is the answer.
Your Next Contract Renewal Is the Decision Point. Ensure You’re Negotiating the Right Terms.
What You’re Actually Choosing Between
Per-seat SaaS gives you access to a platform and assigns you the performance risk. Outcome-based pricing shifts some of that risk back to the vendor but leaves you without the governance structures to push automation beyond 30–50% of its potential. Continuous partnership is the only model in which the vendor’s commercial position and your operational trajectory are pointed at the same outcome.
That last model is harder to negotiate, more demanding to govern, and more likely to produce a vendor who challenges your automation targets rather than manages your satisfaction scores. That friction is the point. A vendor with no financial exposure to your outcomes has no commercial reason to create it.
Return to the three questions at the beginning.
Pricing model. Benchmark access. Margin at risk. If any one of those three answers is unsatisfactory, the decision framework above already tells you which model you are being sold—and what it will cost you over five years.
References:
- 1. 2025 McKinsey LIMRA Insurance 360 Benchmark
- 2. RPA Tech
- 3. BCG
- 4. Prosci
- 5. TurnerLabs
- 6. McKinsey
- 7. Deloitte 2026 Global Insurance Outlook
- 8. PwC
- 9. Fabio Faschi
- 10. IBM
- 11. Bain & Company
- 12. 2025 McKinsey LIMRA Insurance 360 Benchmark
