Innovation Funnel: U.S. Air Force

Productable was building an innovation management platform, and the U.S. Air Force was one of its most demanding clients. The Air Force's innovation program had a real problem: it could generate ideas, but it couldn't systematically advance them. The submission process was clear, but the evaluation and advancement process was opaque -- contributors didn't know what happened to their ideas after submission, and program managers didn't have a reliable framework for deciding which ideas to advance.

The result was a pipeline that looked full but wasn't moving. Ideas accumulated at each stage without clear advancement criteria. Contributors who submitted ideas and heard nothing back stopped submitting. Program managers who lacked a structured evaluation framework defaulted to advancing ideas based on familiarity rather than merit. The innovation program was generating activity without generating outcomes.

The tribal silo problem compounded everything. Airmen at different bases were working on identical challenges -- logistics inefficiencies, maintenance workflow gaps, communication breakdowns -- without any mechanism to connect them. The same problem was being solved six times over in six different locations, with no cross-pollination of what worked. The institutional knowledge was there; the connective tissue was missing.

The language problem was equally damaging. When an idea didn't advance, the system called it a failure. "Fail" ends a conversation. It signals to the contributor that their judgment was wrong, their effort was wasted, and future submissions carry the same risk. The psychological safety required for innovation -- the willingness to propose something that might not work -- was being systematically destroyed by a single word.

My insight was to treat the innovation pipeline as a decision system, and to redesign both the language and the architecture simultaneously. I replaced "Pass/Fail" with "Graduate/Remediate": Graduate means the idea advances; Remediate means the idea needs more development, and here is exactly what it needs. Remediate invites collaboration. It tells the contributor: your idea has merit, and I want to help you make it stronger. This single language change had a measurable impact on re-submission rates.

The Echo System addressed the tribal silo problem. By connecting Airmen in similar roles across disparate bases -- creating structured channels for cross-pollinated knowledge sharing -- I built the connective tissue that transformed isolated problem-solvers into a networked intelligence ecosystem. The same insight that solved a logistics problem at one base could now be surfaced to the Airman facing the same challenge at another base within days rather than years.

But the most instructive lessons from this engagement came from the feedback research -- and they were not what anyone expected.

Leadership told us they wanted more submissions. We took that at face value and optimized for volume. Submission rates increased. Then the real problem surfaced: leadership was overwhelmed. They didn't have the bandwidth to review the increased volume with the depth and quality that would make feedback meaningful to contributors. The ask for "more submissions" was actually an ask for a better signal-to-noise ratio -- for a pipeline that surfaced the right ideas, not more ideas. The lesson: when a stakeholder asks for more of something, the real requirement is almost always better quality of that thing. Volume without evaluation capacity is not a feature; it is a liability.

The contributor feedback was equally revealing. Airmen told us they needed a mobile application -- they were in the field, not at a desk, and the submission experience was built for a desktop environment that didn't match their operational reality. They needed audible capture: the ability to dictate ideas and observations in the moment, before the insight was lost. They needed an AI compilation engine that could take their raw voice notes and transform them into structured, reviewable submissions with feedback, insights, and guidance built in. And they were managing their innovations across too many tools: multiple entry points, fragmented tracking, no single surface where they could see the status of everything they had submitted or were working on.

We ran several design exercises to address these needs. They all failed. The reason was not execution -- it was scope. We were designing in a vacuum, assuming Productable was the only tool in the ecosystem. We hadn't addressed the elephant in the room: how and where did our platform fit and play with the other innovation management platforms already in use? The Air Force's innovation ecosystem was not a blank slate. There were existing tools, existing workflows, and existing integrations. Until we answered the ecosystem positioning question -- what is Productable's specific role, and what does it hand off to adjacent platforms -- every feature we built was solving the wrong problem.

The ecosystem positioning work became the foundation for the roadmap. Once we defined Productable's lane -- the submission, evaluation, and advancement layer -- we could design the mobile capture and AI compilation features as inputs to that lane, not as replacements for the broader ecosystem. The tool fragmentation problem was solved not by building everything, but by defining the handoffs clearly.