Pilot hit its targets but no contract? Engage the budget owner

Why does a pilot that hit its targets stall?

A pilot that hit its targets stalls because it was paid from an experiment budget. Production is paid from a run budget, and that budget’s owner never took part. A run budget pays for a system year after year. It usually belongs to the business line that books the benefit. An innovation team can fund an experiment. It can’t commit a business line to licence fees, integration and support every year, and that commitment is what a production contract is.

So the pilot answered its sponsor’s question: does the product work? Production asks other questions, and other people answer them. The run-budget owner asks whether the result is worth a line in their budget. Finance asks whether the benefit is real money. IT security and risk ask whether the system can go live. Procurement asks how the contract gets signed. None of them was in the pilot, so none of them has a reason to say yes.

The symptoms are familiar: a warm readout with no decision date, a request for a second pilot in another unit and a close date that moves every month.

Those seats belong in the deal early. Ebsta and Pavilion’s benchmark of 655,000 opportunities found win rates rose by 55% when decision makers were actively involved in the first 2 stages of the sales process.^↗ Gong’s analysis of 1.8M deals found multi-threading lifted win rates by 130% on average in deals over $50k.^↗ A pilot with 1 engaged sponsor is a single-threaded deal, however good its results.

Write the production seats into the pilot plan

Plan the production decision inside the pilot. Name each seat it needs, give each seat an owner on your side and a date, then agree the plan with the sponsor before kick-off. A plan with names and dates can be checked each week; “we’ll loop people in” can’t.

Start with the run-budget owner. That is the leader whose budget will carry the running cost and whose numbers move if the product works. Meet them, or the person who reports to them, before the pilot starts. Agree 3 things: the measure the pilot is judged on (in their terms), the production price and the date they’ll decide.^↗ A pilot judged only on the sponsor’s measure gives the budget owner nothing to approve.

If the sponsor can’t name the run-budget owner before kick-off, treat that as your first finding. Ask who funded the last system that went from pilot to production there, and start with that person.

The other seats follow the pilot’s calendar. IT security and risk get the questionnaire and architecture documents in week 1, so their review runs alongside the pilot. Finance agrees by mid-pilot how results become money: hours released, losses avoided or revenue won. Procurement confirms the contract route before the readout. The readout itself goes to the budget owner, with the sponsor presenting.^↗

Match each seat with someone on your side who can answer its first question. Your CEO or a senior leader meets the budget owner, your security lead meets IT security and a value lead who can read a P&L meets finance. If a seat’s date passes without a two-way exchange, it opens the next pilot review and gets a new date and a reason.

Pilot success alone should never count as pipeline

We think a pilot that has engaged only its sponsor has found a supporter and should stay out of the pipeline. Our rule: the production deal becomes a qualified opportunity when at least 2 seats are engaged, including the run-budget owner or their direct report.^↗ The direct report counts because budget owners at large companies often hand the evaluation to someone who carries it into the budget round. Engaged means a two-way exchange with that person in the last 90 days: a call, a meeting or a written reply to a question you asked.^↗ A forwarded deck or the sponsor’s account of a meeting doesn’t count.

We hold that line for 3 reasons. First, a single engaged person isn’t a qualified opportunity while the rest of the buying group hasn’t been reached. A sponsor without a budget is exactly that person. Second, multi-threading happens reliably only when a plan names who talks to whom and by when. Left to a busy sponsor, the introduction tends to come after the readout, when the pilot’s news is already old. Third, a forecast that counts sponsor-only pilots rests on stage definitions nobody can check, which is how a healthy-looking pipeline still misses.

From there the deal moves like any other. It leaves evaluation once IT security is engaged too, and enters commit only with every required seat engaged and a close date inside the quarter.^↗ None of this makes the contract certain. It tells you early whether there is a buyer. It also gives you a measure to watch across your pilots: the share with the run-budget owner engaged.^↗

How do you fix a pilot that already ended?

Treat it as a new deal with 1 engaged seat, and reach the budget owner through your sponsor. Start by calling the pilot complete. Record what it proved and take the production deal out of pipeline until it passes the 2-seat rule.^↗ Then work the steps below, each with an owner on your side and a date.

The introduction should help the sponsor, because their pilot only becomes a result when production funds it. Ask for 1 meeting with the budget owner or the person who reports to them, and say what it is for. You want to show what the pilot proved in that owner’s measure, with the production price and a proposed decision date. Send the person from your side who matches the owner’s level.

Translate the result before that meeting. Model accuracy or minutes saved per case prove the product works. The owner needs the line they answer for, such as cost per case, cycle time, losses or revenue. Finance should agree that translation first.

Then offer a decision the owner can make: a paid first phase that fits this year’s run budget, or a dated line in next year’s plan. If nobody with a run budget will meet you, close the production opportunity with that reason. Set a date to ask again before the next budget round. Each free extension of a finished pilot spends your engineers’ time and moves no decision.

How does the seat plan run at an insurer?

It runs the same way, with 1 added step: an AI governance review. At a carrier the innovation team typically holds pilot money and no production budget. The chief claims officer usually holds the budget for claims AI, and model risk and the CISO can each stop the deal.^↗ Under the NAIC model bulletin, an insurer’s AI programme should cover systems bought from vendors, with due diligence and, where appropriate, audit rights in the contract.^↗ Senior management runs it, accountable to the board.^↗ Open that review in week 1 of the pilot.

In BCG’s survey, only 7% of insurers had brought their AI systems to scale, and about 2 in 3 were still piloting.^↗ In a survey of insurtech founders and senior leaders, 48% of responses rated carrier partnerships low on budget and funding, against 22% on executive sponsorship.^↗

Take an illustrative case: a fictional claims AI vendor whose 8-week pilot at a fictional top-20 P&C carrier hit every target.^↗ After 6 months with no contract, its €350k-a-year production deal has 1 of 5 required seats engaged: the innovation lead.^↗ The claims chief’s side, claims finance, the CISO’s team and model risk are silent, so the deal leaves the forecast. In week 1 the innovation lead introduces the VP of claims operations, who reports to the chief claims officer.^↗ In week 2 the vendor’s CEO meets the VP with the results in claims cycle time and adjuster hours.^↗ That makes 2 of 5 seats, and a qualified opportunity again.^↗ Security documents go to the CISO’s team that week, and claims finance agrees the benefit in week 4.^↗ By week 6 all 5 seats are engaged, with a decision date in the claims budget round.^↗

When a board pushes AI into production, the decision moves to the executive who will run it. For underwriting AI that is usually the chief underwriting officer.^↗ Engage that executive, or the COO where underwriting operations report there, with the innovation team as your champion.