The CEO Who Admits AI Isn't the Answer: What InsTech Leaders Know That Others Don't
The loudest voices on AI are almost never the ones actually making it work. The InsTech leaders worth listening to are the ones who will look you in the eye and tell you AI is not the answer to your problems.
That is not pessimism. That is the most commercially useful thing a senior operator can say right now, and Irish business should be paying close attention to why they say it.
The Gap Between the Headlines and the Boardroom
Insurance is one of the most data-saturated industries on earth. Every claim, every renewal, every fraud pattern, every weather event sits in a database somewhere. If any sector was going to be transformed overnight by AI, insurance was first in line. The headlines said so. The conference circuit said so. The vendor decks definitely said so.
What actually happened was more instructive. A string of InsTech CEOs, the ones running real books of business rather than pitching seed rounds, started drawing a line between AI as a product and AI as an enabler. The distinction sounds like wordplay until you see the numbers behind it.
Lemonade, the AI-first American insurer that was supposed to rewrite the industry, posted a combined ratio above 90% for years running. A combined ratio above 100% means you are paying out more in claims and costs than you collect in premiums. The market rewarded the story. The story and the P&L were two different documents. Meanwhile, incumbents who quietly wired AI into their underwriting workflows, not as a headline, but as a process improvement, began compressing their loss ratios by three to five points. Three points on a large book is not a rounding error. On a €500 million premium book, that is €15 million finding its way back to profit.
What 'Enablement' Actually Means in Practice
The leaders getting this right are not replacing underwriters with algorithms. They are giving underwriters better information faster. That is the whole framework, and it fits on one page.
Step 1: Map the decision, not the department. Before any technology conversation, you identify exactly which decisions in your business carry the most financial weight. In insurance, it is the pricing decision at point of quote. Everything else is downstream of that moment.
Step 2: Ask what information would change the decision. Not what data you have. What data, if you had it at the moment of decision, would produce a better outcome. This is producer thinking, not consumer thinking. You are not buying a platform and hoping for the best.
Step 3: Build the pipe before the pump. Most Irish organisations that have struggled with digital transformation spent money on the AI tool before they had clean, connected data to run through it. The AI is the pump. If the pipe is broken, the pump is irrelevant. This is where the majority of wasted budget goes.
Step 4: Measure the decision, not the technology. The KPI is not AI adoption rate or queries processed. The KPI is whether the underwriter made a more accurate pricing call this quarter than last quarter. Technology is a tool, not an asset, and you do not report tool usage to your board.
The order matters because each step is a gate. Organisations that jump to Step 4 without doing Step 2 spend two years measuring the wrong thing and wonder why the transformation stalled.
The Irish Context
Irish insurers carry a specific structural challenge that makes this conversation more urgent than it might appear. Motor and liability premiums in this country remain among the highest in Europe, not primarily because of fraud, but because of claims inflation and an underwriting market that historically lacked the pricing granularity its British and continental peers built years ago. The Personal Injuries Guidelines introduced in 2021 were supposed to correct the claims side. They helped. But the pricing side still relies on underwriting judgements that are slower and coarser than they need to be.
A mid-sized Irish insurer that can compress its quote-to-bind cycle from four days to four hours, using AI to pull in third-party data on property risk, fleet history, or claims patterns at point of quote, does not just save internal cost. It wins business that the slower competitor cannot even respond to in time. Speed is scarcity. Scarcity is margin. This is the opportunity the InsTech leaders are actually talking about when they say AI is an enabler rather than a product.
What the Honest CEO Sounds Like
The InsTech leader worth your time is not the one on stage promising autonomous claims. They are the one in the meeting saying: our frontline staff make 400 pricing decisions a day, and I want each of those decisions to be 5% more accurate by end of year. Full stop.
That framing is unglamorous. It does not generate a TED talk. It generates a business that compounds quietly. The employee sees a better tool and does a better job. The customer gets a price that reflects their actual risk rather than a broad actuarial bracket. The company retains more of its premium as profit because the loss ratio tightens. Nobody writes a feature about it.
This is the same logic that explains why Irish financial firms are rethinking security from the inside out rather than buying a perimeter product and calling it done. The underlying principle is identical. Better information at the point of decision beats a better-looking system at the point of sale.
The Posture That Wins
Entrepreneur versus employee. Producer versus consumer. These binaries matter here because the boardrooms that are winning are producing outcomes, not consuming platforms. They bought a tool, built the process around it, and measured what the process changed. The boardrooms that are losing bought the platform, announced the transformation, and are now quietly trying to explain to their CFO why the cost went up and the benefit has not arrived yet.
AI is not the answer. Better decisions are the answer. AI is one way to get there, when it is wired into the right moment, with the right data, in the hands of someone who still understands the risk.
That is the masterclass. It is not on a stage. It is in the accounts.
