A city agency pilot program is a structured, time-limited test of your technology inside a municipal department, usually run by the city or by a partner nonprofit, after which the agency decides whether to formally procure what you built.
Getting into one is mostly preparation and framing. Cities cannot buy unproven technology outright, so the pilot is the door in, and the people who open it care far more about a measurable operational problem than about your product.
The process below is what I would hand a founder on day one. It takes a few weeks of prep, not months, before you ever send an email.
Table of Contents
- What You Need Before You Contact Anyone
- Step-by-Step: How to Get a Pilot Program with a City Agency
- 1. Identify a Specific City Problem
- 2. Build Evidence That the Problem Exists
- 3. Create a Minimum Viable Pilot
- 4. Find the Right Agency Owner
- 5. Pitch a Low-Risk Collaboration
- 6. Agree on Success Measures and Guardrails
- 7. Get Written Approval and a Next Step
- Common Mistakes When Approaching a City Agency
- Treating a Meeting as Approval
- Pitching a Product When the Agency Framed a Capability
- Ignoring Procurement, Security and Privacy Early
- Using Vanity Metrics
- Leaving Agency Responsibilities Undefined
- Approaching a Department With No Budget and No Owner
- Ignoring the Third-Adopter Strategy
- Frequently Asked Questions
- Does a city pilot program lead to a contract?
- Do I need an existing government contract to get a first pilot?
- How long does a city agency pilot take?
- How do I find city pilot programs in my city?
- Who pays for a city agency pilot?
- What is a pilot in affordable housing?
- Conclusion: Start With One Problem and One Meeting
What You Need Before You Contact Anyone
You need seven things ready, and the first one is not a deck. A clearly defined public problem comes first, because that is what an agency evaluates you against.
- A specific operational problem owned by one department, expressed in the agency’s own words.
- A working prototype or concept, not a slide describing one. A demo that runs beats a roadmap every time.
- Two or three measurable outcomes and the baseline number each one would move.
- A data and privacy position: what you need, what you do not need, and how you would protect it.
- Implementation capacity, including the staff hours, support model and timeline you can actually commit.
- Budget assumptions, covering who pays, what a pilot costs the agency in staff time, and what happens after.
- A one-page pilot proposal that an executive assistant could forward without asking you questions.
The one-page document is the single highest-leverage artifact in this whole process. Programs like Philadelphia’s Pitch & Pilot ran on exactly this kind of short submission, and the NYC DEP Environmental Tech Lab used a defined application window with an information session before it.
Step-by-Step: How to Get a Pilot Program with a City Agency
Seven steps, in order. Skip ahead and you lose weeks.
1. Identify a Specific City Problem
Replace your broad innovation idea with a narrow operational problem that one named agency owns. “We use AI for cities” is not a problem; “field inspectors lose track of overdue sewer inspections because status updates happen in three separate spreadsheets” is.
Look for the same shapes of trouble repeatedly: missed transit inspections, permit-status calls flooding a call center, energy bills nobody can explain to a tenant. Work permit and licensing, permitting and inspections, 311 and constituent services, transit and fleet operations, water and wastewater, workforce development.
One agency, one workflow. If your idea touches four departments, you are four conversations, and none of them will sponsor you.
2. Build Evidence That the Problem Exists
Combine public data, interviews and frontline observation into an evidence base the agency cannot dismiss as vendor hype. Public dashboards, budget documents, open data portals, inspection reports and adopted plans are all fair game and all free.
Then go talk to the people who live the problem, even informally. A rideshare driver, a permitting clerk at a counter, a code enforcement officer. Those conversations tell you whether your framing matches reality, and they usually rewrite your first draft.
Do not overstate. If you want restricted internal data, say so plainly and explain how you would protect it rather than implying you already have access.
3. Create a Minimum Viable Pilot
Define the smallest useful test: which users, which workflow, how long, what it takes from the agency, and what is explicitly out of scope. Small is a feature here. A four-month test on one team is easier to approve than a citywide rollout.
Write down what the pilot will not include. That sentence does more to win a public-sector audience than any product claim, because agency staff are the ones who get asked about scope after you leave the room.
4. Find the Right Agency Owner
Map the problem to the responsible department, the program office, the innovation or transformation team, the procurement contact, and the operational sponsor who would actually use what you built. These are usually five different people, and only one of them can say yes.
Research the contact before you email. Read the department’s annual report, its meeting minutes, its press releases. Note who talks about the theme you are pitching. Then ask for a short introductory conversation rather than a demo, and say in your first line what problem you want to discuss.
The fastest route into a pilot usually runs through an intermediary: a nonprofit business group, a university tech transfer office, a public-private partnership office, or a civic accelerator. Many cities publish calls for applications through those partners rather than on their own sites.
5. Pitch a Low-Risk Collaboration
Position the proposal as a measurable experiment, not a system replacement. Agencies get more proposals than they can evaluate, and the ones that read as a soft pitch lose to the ones that read as a controlled test with an exit plan.
Your one-pager should cover six things: the problem in one paragraph, your approach in two, the timeline, who does what, your privacy and security safeguards, and the result the agency gets. Keep it under a page and free of jargon.
Say what the pilot is not, too. Not a procurement commitment, not a replacement for an existing system, not a transfer of operational control.
6. Agree on Success Measures and Guardrails
Agree on metrics before the pilot starts, not after the results look ambiguous. Specify the baseline, one primary metric, one or two secondary metrics, how they get reported, who owns the data you produce, and what happens if the pilot misses.
Guardrails matter as much as metrics: what data you can and cannot touch, what happens to resident information, what the agency needs to sign off before anything goes live, and the security review you are prepared to pass.
Name the end condition for both sides. A pilot with no defined stopping point turns into an unfunded ongoing favor, which is how relationships get damaged.
7. Get Written Approval and a Next Step
Convert informal interest into one documented action: a discovery call, a letter of support, a pilot charter, a memorandum of understanding, a written test plan, or a procurement review.
A meeting is not approval. If nothing is written down, there is no pilot, and a friendly conversation months from now will not turn into a signature. Ask for the specific artifact that matches where you both are, and agree on who sends it and by when.
Then confirm the fiscal calendar. Municipal money moves on budget cycles, and a pilot approved after a budget is committed can sit idle for most of a year.
Common Mistakes When Approaching a City Agency
Most stalled pilots fail on the same handful of things, and every one has a straightforward fix.
Treating a Meeting as Approval
An enthusiastic conversation is not a commitment. Fix: end every meeting by agreeing on one written next step and a date. If nobody will write anything down, you have an introduction, not a pilot.
Pitching a Product When the Agency Framed a Capability
Agencies describe what they need to do, not what they need to buy. A department asking for better permit-status visibility is not shopping for a permit platform. Fix: restate their goal in their language first, then say what you built and why it fits.
Ignoring Procurement, Security and Privacy Early
Local purchasing thresholds differ by city and state, and so do cybersecurity review, insurance requirements and data-sharing terms. A pilot can die in month four because nobody checked them. Fix: find the contracting officer or procurement analyst in month one and ask what review applies before you promise anything.
Using Vanity Metrics
Sign-ups, page views and press mentions tell an operations team nothing. Fix: pick a number tied to staff time, response time, backlog or cost, and show the baseline you measured before the pilot.
Leaving Agency Responsibilities Undefined
Vendors routinely assume the city will assign staff, extract data and schedule interviews. Fix: list every input the agency must provide, and mark clearly which ones need a name attached to them.
Approaching a Department With No Budget and No Owner
Not every department runs pilots, and plenty have no discretionary budget line for experiments. Fix: ask early whether the agency has an innovation office, a funded program, or a partner nonprofit running a cohort. If none of those exist, your target city may be your second or third pilot, not your first.
Ignoring the Third-Adopter Strategy
The best first customer is a city that already runs pilot programs and has adopted something similar before. Being first is expensive: you inherit every problem the agency has and no precedent to lean on. Fix: target a third adopter, prove yourself there, then use the reference to reach the city you actually want.
Frequently Asked Questions
Does a city pilot program lead to a contract?
No, not automatically. A pilot is a time-limited test, and a separate procurement decision still has to run afterward. The NYC DEP Environmental Tech Lab was built on exactly this design: applicants work with departmental staff on a pilot, then the agency decides whether to recommend adopting the technology through normal purchasing. Design your pilot so the results and usage data are procurement-ready.
Do I need an existing government contract to get a first pilot?
No. Most programs run by city innovation offices and partner nonprofits are built for vendors with no government track record at all, because the whole point is to test something unproven. What matters more than past contracts is having a working product, a clearly scoped problem and enough implementation capacity to finish the pilot properly.
How long does a city agency pilot take?
Plan for four to six months of actual testing, plus one to two months before that for security review, procurement paperwork and data agreements. Cities move slowly on paper and slowly on schedules, so build the calendar around the agency’s budget cycle rather than your product roadmap. If a program publishes a call for applications with deadlines, work backward from that date.
How do I find city pilot programs in my city?
Search your city domain for pilot programs, innovation lab and accelerator terms, and check the economic development or public-private partnership office site. Search Grants.gov and your state’s open data portal for grant-funded cohorts. Then check whether a local nonprofit business group or university runs a pilot cohort, since many cities accept applications through a partner rather than directly.
Who pays for a city agency pilot?
It depends on the model. Some programs fund the pilot and cover a vendor stipend. Some cities fund only the staff time they contribute, while the vendor provides the technology. Others are grant-funded, where federal or state money pays for the evaluation and the vendor absorbs part of the cost. Ask early, because who pays changes the scope more than anything else.
What is a pilot in affordable housing?
In housing departments, a pilot usually tests a specific intervention in a limited set of units or households: faster permit and inspection turnaround for a rehabilitation program, a benefits screening step for residents at risk of eviction, or a reporting tool that tracks code violations. The same rules apply as elsewhere. One housing authority, one bounded test, agreed metrics, and a defined end point.
Conclusion: Start With One Problem and One Meeting
Pick one specific city problem, write the one-page brief that frames it, and ask the responsible agency owner for a short problem-validation conversation rather than a demo. That single week of preparation puts you ahead of most vendors in the queue.
Keep the pilot small, the metrics agreed, and the paperwork started early, because the pilot itself is the evidence that makes the contract conversation possible later.
