by Braden Kelley and Chateau G Pato
What Should You Ask Before Funding an Innovation Pilot? (Short Answer)
Before funding any innovation pilot, ask whether you have a named human problem, a clear beneficiary, a behavior-based success measure (not theater metrics), a durable owner after applause, decision rights to change something real, kill criteria, a path beyond the A-team, a plan to retire the old way, clarity on whose interest is optimized, guardrails against a hard landing, and a useful “no.” Most wasted pilot spend fails at the funding gate — not the tech gate.
The eleven questions: (1) What human problem are we funding? (2) Whose life gets better? (3) What behavior will we measure? (4) Who owns this after applause? (5) What can we change if we learn something uncomfortable? (6) What is the kill criteria? (7) How does this scale beyond the A-team? (8) What old way will we turn off? (9) Whose interest is optimized? (10) What hard landing are we refusing? (11) If we say no, what do we still owe?
The Pilot Is a Decision, Not a Hobby
I have watched organizations fund pilots the way people buy gym memberships in January — with optimism, a logo on the slide, and almost no plan for Tuesday in March. Then they act surprised when the demo works and the operating model does not move.
Human-centered innovators treat the funding meeting as the first design decision. A pilot is not a hobby with a budget code. It is a bet that should either earn the right to change how people work — or be stopped cleanly so capacity returns to a better problem.
If you cannot answer the eleven questions below, you are not being agile. You are underwriting theater.
| Question | Red-flag answer | Fundable answer |
|---|---|---|
| 1. Human problem | “Explore AI / innovate” | One sentence: person + stuck job |
| 2. Beneficiary | “The business” | Named role + felt success |
| 3. Success measure | Demos, idea count | Behavior change you can observe |
| 4. Owner after applause | “Innovation team / vendor” | BAU workflow owner with rights |
| 5. Decision rights | Cannot touch policy or process | Pre-agreed levers to move |
| 6. Kill criteria | “We’ll keep learning” | Time box + fail threshold + stopper |
| 7. Beyond A-team | Heroes + manual glue | Median user + debt named |
| 8. Turn off old way | “Run both forever” | Retire date or condition |
| 9. Whose interest | Deflection / savings only | Human success + economics |
| 10. Hard landing | No failure mode | Named harms + guardrails |
| 11. Value of a no | Only yes-path thinking | Learning returned; next bet clearer |
1. What Human Problem Are We Actually Funding — in One Sentence?
Why it matters: Pilots without a named human job become feature tourism. The slide says “innovation.” The calendar says “wandering.”
Red-flag answer: “Explore generative AI.” “Test the platform.” “Build our innovation muscle.”
Fundable answer: One sentence with a person, a moment, and a stuck job — e.g., “First-line agents cannot resolve billing exceptions without three systems and a supervisor.” If you need a paragraph, you do not have a problem yet. You have a vibe.
2. Whose Life Gets Better If This Works — Customer, Employee, or Both?
Why it matters: Orphan beneficiaries become orphan adoption. “Efficiency” without a face is how pilots optimize the spreadsheet and punish the humans.
Red-flag answer: “The business.” “Shareholders.” “Everyone, eventually.”
Fundable answer: A named role and what success should feel like — less shame, less rework, more confidence, more time for judgment. If both customer and employee gain, say how. If only one gains at the other’s expense, say that too. Honesty is cheaper than a failed rollout.
3. What Will We Measure That Is Behavior — Not Theater?
Why it matters: Activity metrics fund costume. Ideas generated, demos held, patents filed, “innovation NPS” — applause instruments, not proof.
Red-flag answer: Engagement with the pilot. Number of prompts. Workshop attendance.
Fundable answer: Observable human behavior on a journey — completion in one attempt, abandoned workaround, lower effort, faster time-to-confidence, fewer repeat contacts, retained customers, managers using the new path without shadow process. If you cannot name the behavior that changes when you “win,” do not fund the costume.
4. Who Owns This After the Pilot Applause Ends?
Why it matters: No durable operator is how you buy pilot purgatory. Projects end. Vendors leave. BAU inherits a half-changed Tuesday.
Red-flag answer: “The innovation lab.” “The vendor.” “We’ll figure out ownership at scale.”
Fundable answer: A named workflow owner in the business with decision rights — someone whose job still exists when the ribbon-cutting is over. If that person is not in the funding meeting, you are funding an exhibit.
5. What Are We Empowered to Change If We Learn Something Uncomfortable?
Why it matters: Methods without mandate are innovation cosplay. Sticky notes cannot move policy, incentives, or process they are forbidden to touch.
Red-flag answer: “We’re just testing technology.” “No process changes in phase one.”
Fundable answer: Pre-agreed levers — which policy, metric, handoff, or script can move if evidence says the current way is the problem. Learning that cannot change anything is entertainment with a SOW.
6. What Is the Kill Criteria — and Who Has the Courage to Use It?
Why it matters: Endless “we’re learning” is luxury consumption. Without stop-rules, pilots become pets.
Red-flag answer: No end date. No fail threshold. “The steering committee will decide later.”
Fundable answer: A time box, an evidence threshold, and a named person authorized to stop. Killing a weak pilot is not anti-innovation. It is how you protect the next good bet.
7. How Does This Scale Beyond the A-Team — or Are We Funding Heroics?
Why it matters: Hand-picked enthusiasts, clean data, and spreadsheet glue prove almost nothing about the median Tuesday.
Red-flag answer: “We’ll harden it later.” “Our best people will run it first.”
Fundable answer: Design assumptions for the median user and median manager. Integration and data debt named as costs in the pilot ask — not as a surprise after the applause. If only heroes can run it, you are funding burnout with a demo.
8. What Old Way Will We Turn Off If This Works?
Why it matters: Shadow paths keep adoption optional. Dual operating systems drain trust and attention.
Red-flag answer: “We’ll run both for a while” with no end condition.
Fundable answer: An explicit retire date or trigger for the workaround, spreadsheet, or legacy path. If you cannot turn something off, you have not finished designing the bet — only the brochure.
9. Whose Interest Is This Optimizing — Human Outcome or Cost-Containment Theater?
Why it matters: Pilots that succeed only on deflection, speed, or savings often fail on trust. Customers and employees keep score in their bodies.
Red-flag answer: Success equals containment, handle time, or “AI adoption” alone.
Fundable answer: Human success and economics in the same sentence — resolve completely, reduce effort, preserve dignity, and improve cost-to-serve or revenue quality. If the pilot cannot say who is optimized, assume it is the slide.
10. What Hard Landing Are We Refusing — and How Do We Prevent It?
Why it matters: Soft landings are designed at the funding gate. Hard landings are what you get when efficiency is the only value on the dashboard.
Red-flag answer: No failure mode for customers or employees. “We’ll monitor.”
Fundable answer: Named harms you refuse — loop traps, denser busywork, surveillance personalization, blocked escalation, rubber-stamp humans — plus guardrails: undo, consent, handoff, protected deep work, decision rights. If you cannot describe the hard landing, you are not ready to fund the soft one.
11. If We Say No, What Do We Still Owe the Organization?
Why it matters: A good no is innovation hygiene. Fear of looking “anti-innovation” is how bad pilots get yeses.
Red-flag answer: Only a yes-path. Silence after rejection. Political punishment for stopping.
Fundable answer: Even a no returns something — documented learning, clarified problem, freed capacity, a sharper next bet. Leadership that can only fund yeses will eventually fund theater.
How Do You Decide Whether to Fund an Innovation Pilot?
Run the eleven questions in one meeting. Score each green, amber, or red.
- Three or more reds: Do not fund. Fix the ask or kill it.
- Ambers: Fund only with written conditions and a named owner for each condition.
- All green or green-with-conditions: Fund the smallest bet that can falsify the idea — not the largest deck that can impress the room.
Funding a pilot without these answers is not agility. It is underwriting theater. Real innovators do not fear the questions. They fear spending a year proving something that was never designed to land with humans in the first place.
Frequently Asked Questions
What questions should you ask before funding an innovation pilot?
Ask what human problem you are funding, whose life gets better, what behavior you will measure, who owns the work after applause, what you can change if you learn something uncomfortable, what the kill criteria are, how it scales beyond the A-team, what old way you will turn off, whose interest is optimized, what hard landing you refuse, and what you still owe the organization if you say no.
How do you decide whether to fund an innovation pilot?
Score the eleven funding questions green, amber, or red in one meeting. Three or more reds means do not fund. Ambers require written conditions. Fund only the smallest bet that can falsify the idea when answers are green or conditionally green.
What is kill criteria for an innovation pilot?
Kill criteria are the pre-agreed time box, evidence threshold, and authority to stop a pilot that is not earning the right to continue. Without them, “we’re learning” becomes an excuse to keep funding pets instead of progress.
Why do innovation pilots waste money?
Most wasted pilot spend fails at the funding gate: no named human problem, vanity success metrics, no durable owner, no decision rights, no kill criteria, A-team heroics, dual operating systems, and optimization for slides or containment rather than human outcomes.
Should you fund an innovation pilot without a scale plan?
Not if “no scale plan” means no owner, no path beyond heroics, and no intent to retire the old way. You can fund a small learning bet — but only with kill criteria and clarity about what operating-model questions the pilot must answer before more money arrives.
Image credits: Google Gemini
Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Google Gemini and Cursor to clean up the article, add images and create infographics.
Sign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.