.png)
The Pilot Went Great. So Why Isn't Anyone Buying?
From pilot to procurement: how public safety technology companies turn a good evaluation into a funded deployment.
Too many emerging technologies for public safety has some version of this story.
The pilot went well. The officers or firefighters who used it liked it. The chief said kind things at the closeout meeting, and on video! Someone even asked how many more licenses they could get.
Then six months went by and crickets.
No contract. No line in the budget. The champion got promoted, or retired, or got pulled onto a CAD migration. The pilot didn't fail. It just never turned into anything.
With IACP wrapping up in Orlando this week, a lot of vendors are flying home with a notebook full of "let's set up a pilot." Before any of those start, it's worth asking a blunt question: are we testing whether the product works, or whether this agency can actually adopt it?
Those aren't the same thing. Most stalled pilots were built to answer the first question when the buyer needed an answer to the second.
Product pilots and adoption pilots
A product pilot asks whether the technology does what the vendor says. Does the drone launch on time, does the transcription hold up, does the map refresh.
An adoption pilot asks whether this agency can use it every day, support it, pay for it next year and buy it through its own process.
The first question has one stakeholder: the people using the product. The second has six or seven, and most of them never touch the product during the pilot. That's where things fall apart. The vendor proves the software works. The agency learns its people like it. Nobody has worked out who pays for year two.
Key Insight: The purpose of a pilot is to inform an adoption decision. If nobody agreed up front what that decision is, who makes it and when, the pilot has nothing to inform.
AI is the sharpest example right now, but not the only one
AI makes this gap easy to see because use is climbing faster than the systems around it. In the 2026 Verizon Frontline survey, 32 percent of public safety respondents said they use AI daily at work, up from about 11 percent in 2025. The same survey found 61 percent citing budget constraints, 66 percent ranking unreliable networks among their top three concerns, and 44 percent pointing to a lack of interoperability (GovTech). People are using the tools. The budgets, networks and integrations needed to make them permanent are lagging behind.
The OECD sees the same pattern at the national level. Its 2026 Digital Government Outlook says:
"Challenges to measuring the impact of AI limit decision-making and contribute to a proliferation of pilots." (OECD, Digital Government Outlook 2026)
The report adds that "only 10 of 36 OECD countries (28%) report any financial or non-financial impact measurement." If national governments struggle to measure results, a 60-officer department or a county emergency management office will too. Emergency management is the clearest case. A set of reports summarized by GovTech found that "many smaller EM offices operate with one full-time staff member or fewer" and that "many agencies lack the capacity and trust needed to use those tools responsibly" (GovTech).
Practitioners trying to get government AI out of the sandbox say much the same. In a GovTech piece by Julia Edinger, George Mason University chief AI officer Amarda Shehu noted that "a lot of organizations underestimate the data maturity needed to support these models." I-Sah Hsieh of North Carolina's Department of Information Technology pointed out that "very few have publicly declared how they decide when to decommission an AI system." South Carolina AI director Rich Heimann described an "asymmetry": AI companies are still learning the obligations of public-sector work, while traditional government vendors are still learning AI.
That asymmetry is exactly what a well-designed pilot should close.
And this isn't only an AI problem. Drone as First Responder programs face the same handoffs. Discussing CPSE's July 2026 report on DFR programs, CPSE board president and Poudre Fire Authority Chief Derek Bergsten said peer departments helped his agency work through "the same challenges we faced," ranging "from airspace coordination to real-time video integration with dispatch" (FireRescue1). None of that shows up in a demo flight. All of it decides whether the program lasts.
Find the real buying committee
The person who loves your product during a pilot usually can't buy it. Here's who typically can, or can stop it:
Stakeholder | What they need to see from the pilot |
End users (officers, firefighters, telecommunicators) | It saves time or improves safety on a real shift |
Operational sponsor (deputy chief, commander) | A measurable change against today's baseline |
IT and security | Integration effort, CJIS and data handling, support load |
Finance and budget office | Total cost over three to five years, and a funding source |
Procurement and legal | A compliant path to purchase and acceptable contract terms |
Agency head and elected officials | Public benefit they can explain and defend |
Federal agencies with full contracting staffs run into the same wall. The GAO's 2026 review of AI acquisitions lists challenges including "traditional acquisition time frames and contract approaches," "early testing and continuous evaluation," and "artificial intelligence pricing and overall cost." Some officials said they "struggled to establish effective evaluation factors for source selection" (GAO-26-107859). Local procurement works differently, but the problems rhyme. If the agency can't score your product in a competitive buy, your pilot results may never make it into the decision.
The FirstWave Pilot-to-Adoption Framework
This is the five-stage approach we use with clients. Each stage ends with something concrete in hand.
Qualify. A named operational problem, an executive sponsor, the person with decision authority and a plausible budget source. No sponsor, no pilot.
Design. A written evaluation plan with today's baseline, three to five success metrics, the evaluation period, and security and integration requirements, signed off by the sponsor.
Validate. Field results, end-user feedback, the real implementation effort and an honest list of limitations. Agencies trust a vendor more when the vendor writes down what didn't work.
Fund. A procurement strategy, a budget source, total cost of ownership and an internal business case the sponsor can carry into budget season.
Deploy. Contracting, onboarding, training, support and a plan to keep measuring after go-live.
Most pilots jump from Design to Validate and stop there. Fund is the stage that turns a good result into a purchase order, and it has to start in week one, not after the closeout meeting.
Six questions to settle before the next pilot starts
What changes operationally if this works, and how will we measure it against today?
Who has authority to make the buying decision, and have they agreed to the success criteria?
Where does the money come from in year one and in year two: operating budget, a grant, a capital plan?
What contract vehicle gets us from pilot to purchase without starting over, such as a cooperative contract or an existing agreement?
Who owns training, integration and support after the pilot team goes home?
What happens the day after the pilot ends if the results are good? And if they're mixed?
If those answers are fuzzy, the pilot probably isn't ready. That's fine. Working through them is the most valuable thing a vendor and an agency can do together before the first device ships.
A note for agency leaders
A pilot is also procurement homework. Setting success criteria in advance protects you as much as the vendor. It gives you a defensible record when you go to council for funding, and a clean reason to walk away if the results don't hold up.
The bottom line
For technology founders entering public safety: before you offer another free pilot, ask whether the agency has a defined path to make a purchasing decision. If it doesn't, help build one. That's the work that turns a pilot into a customer.
FirstWave Strategies helps public safety technology companies design pilots that lead to procurement, build partner strategies and grow into new agency markets. If you have a pilot in flight, or one about to start, let's talk.
Frequently asked questions
What is the difference between a product pilot and an adoption pilot? A product pilot tests whether a technology works. An adoption pilot tests whether a specific agency can use, support, fund and buy it through its own procurement process.
Why do public safety technology pilots stall after a successful evaluation? Usually because success criteria, the decision-maker, the funding source and the contract path weren't agreed before the pilot began. Strong results can't fill those gaps later.
How long should a public safety technology pilot run? Long enough to cover the evaluation period set in the design stage, and timed so results land before the agency's budget deadline.
Footnote: This post was drafted with the help of AI and edited by the author. All quotes and figures link to their original sources.
Comments