Pilot Purgatory

Good evening, reader.

Editor's note: This week's piece is for the operators. It names a pattern most founders selling into Australian institutions know in their bones but rarely see written down: the pilot that concludes, glowingly, and converts into nothing. The argument is that this is not bad luck and it is not culture. It is a market structure, and market structures can be renegotiated.

In today's Innovation Commons:

• Grant assessors across Australia's 2026 commercialisation programs now treat pilot contracts and commercial partnerships as the standard evidence of traction (The Investor Standard).

• The flagship programs, from the $15 billion National Reconstruction Fund to the Industry Growth Program's $50,000 to $250,000 grants for validated prototypes, now form a de facto pathway linking research, proof of concept, pilot activity and commercial growth (GrantHelper; The Investor Standard).

• That pathway has a missing gate. Nothing in the system prices, measures or rewards the conversion of a successful pilot into a procurement contract.

• The result is a proof economy: pilots that clear in the grant market and the institutional reputation market while the revenue market never opens. This piece is about how the loop works, and how an operator steps out of it.

Proof clears in two markets

There is a founder somewhere in Australia this evening preparing a kickoff deck for her third pilot with the same institution. The first, two years ago, was declared a success. So was the second. Each produced an internal report, a conference case study and a warm introduction to another business unit, which proposed a pilot. None produced a contract.

She is not unlucky. She is inside the dominant mode of engagement between Australian institutions and the companies trying to sell to them, and the mode has a shape worth naming: the pilot has stopped being a step toward procurement and has become a substitute for it.

Look at how the 2026 funding landscape prices evidence. The assessment of the current grant cycle is blunt: founders must demonstrate traction, and the accepted proof is revenue milestones, pilot contracts and commercial partnerships. GrantHelper's survey of the year's programs describes the flagship schemes as a pathway running from research through proof of concept and pilot activity to commercial growth. Each filter is sensible in isolation. Together they close a loop. The institution needs pilots to demonstrate it is innovating. The founder needs pilots to demonstrate she is fundable. Both sides can transact indefinitely in proof, because proof now clears in two markets, the institutional reputation market and the grant market, while the one market that sustains a company, revenue, never has to open.

A pilot is cheap optionality for the buyer, and the option premium is paid by the seller. For a modest fee, often for nothing, an institution acquires the right but not the obligation to adopt. In any other market the party granting an option charges for it. In the pilot economy the premium runs the other way: the startup pays it, in engineering time, security reviews, integration work, bespoke reporting, and the executive attention of a ten-person company pointed at a ten-thousand-person one. The option holder pays from an innovation line item that exists precisely so the operational budget never has to be touched.

Why the door never opens

The usual explanation is culture: risk aversion, bureaucratic timidity. That gets the diagnosis backwards. Everyone involved is behaving rationally inside the incentive structure they were handed. Nobody in Australian institutional life has been sacked for a pilot that lapsed. Pilots do not fail; they conclude. A procurement decision has an owner, an audit trail, and a downside that attaches to a named person. The asymmetry is total. The pilot offers upside optics with no accountable downside. Procurement offers accountable downside with almost no personal upside.

The org chart compounds it. Pilot money and procurement money live in different parts of the building, controlled by different people, on different planning cycles. The innovation team that ran the pilot cannot convert it. The operations team that could convert it never sponsored it and inherits none of the credit. The pilot succeeds, the report circulates, and the deal falls into the gap between two budget owners. No amount of product quality bridges that gap, because the product was never the blocker.

What conversion discipline looks like

None of this is an argument against pilots. A well-constructed pilot is the correct instrument for genuine technical uncertainty. The argument is against pilots constructed without an exit, and the fix is contractual, agreed before kickoff rather than negotiated from weakness at day 89.

The elements are unglamorous. Success criteria defined in writing before the pilot starts, wired to a specified procurement pathway: if the criteria are met, the pilot converts through a named mechanism, not into a steering committee discussion. A budget owner from operations, not innovation, signing the pilot agreement, so the person who can buy is implicated from day one. A sunset clause, so the pilot ends on a date and ending without conversion is recorded as a decision with a stated reason rather than dissolving into silence. And a price that reflects what the pilot costs to deliver, because a free pilot signals that the vendor, too, considers this theatre.

Institutions that adopt this discipline discover something uncomfortable: their pilot volume collapses. Most proposed pilots cannot survive the question of which operating budget would pay for the solution if it works. That is the point. The pilots that survive the question are the ones that were ever going to matter.

The operator's counter-position

Founders cannot reform institutional incentive structures. They can stop subsidising them. Three questions, asked before signing rather than after souring, do most of the work.

First: which operating budget would this solution be paid from if the pilot succeeds, and has the owner of that budget seen this agreement? If the answer involves the innovation team finding out later, the pilot has no exit. Second: when did this institution last convert a pilot into a procurement contract, with any vendor? Past conversion behaviour predicts future conversion behaviour far better than present enthusiasm. Third: what happens on the day after the pilot ends? If the honest answer is that next steps would be discussed, the next step has already been decided.

A founder who walks away from a pilot that fails these questions has not lost a customer. She has declined to pay an option premium to a counterparty who was never going to exercise. The capital and the engineering months that would have gone into a third round of proof can go instead to the small number of buyers who enter pilots because they intend to leave them. They exist: in mining, in parts of defence, in the corners of government that have felt real operational pain.

This reading comes with a falsification test. If your institutional counterparty can name the operating budget, the conversion mechanism and the decision date before kickoff, this piece is wrong about them, and you should take the pilot. If they cannot, the piece is right, and the pilot is the product they are buying.

Australia does not lack prepared ground. The National Reconstruction Fund, the state commercialisation programs, a decade of procurement reform reviews: the furrows have been ploughed. What we have built instead is an ever longer shelf of glassware, contained proofs of concept, roots circling, each one lit well enough to photograph for an annual report. The seedlings are fine. The problem is that someone has to decide to plant one, and we have engineered a system in which that decision is the only act that carries a cost.

— The Editor

P.S. Share this on LinkedIn/X to spark conversations with your colleagues.