Additional Meta Reflection: Access, Membranes & Scale
Beyond the specific process and round design learnings outlined above, our experiences have surfaced a deeper, structural tension that becomes increasingly important as Localism Fund scales:
how to remain open and discoverable to under-networked local actors while also preserving review rigor and staying within realistic operator capacity?
In traditional grant-making, this challenge is well known. As application volume outpaces funding and review capacity, programs tend to drift toward one (or more) familiar failure modes:
- Becoming closed — shifting to invite-only networks or informal insider access to control volume.
- Becoming bureaucratic — adding layers of forms, compliance checks, and procedural hoops that raise barriers for applicants and operators alike.
- Becoming shallow — relying on fast scoring or surface-level review that disproportionately rewards polished narratives over real capacity or evidence.
What makes this tension even more acute today is a shift in the signal environment. AI-assisted grant writing increases the length, baseline polish and apparent coherence of applications, reducing the informational value of narrative quality as an indicator of real capability. Under the pressure of scale, this amplifies the risk of shallow review (where polish is mistaken for substance) and bureaucratic over-correction (where more hoops are added to compensate), while simultaneously increasing the temptation to close open access altogether.
Rather than defaulting toward closure, procedural overload, or shallow review, this tension can could potentially be addressed through explicit design choices and innovation:
- Stage-gate review effort — use enhanced eligibility and fit screening before deep evaluation.
- Shift from narrative to evidence — prioritize verifiable signals (e.g. links to prior outputs) over polished writing.
- Design for capacity — align intake size, review depth, and allocation models with real operator bandwidth.
- Use AI as assistive infrastructure — apply carefully constrained AI to support early screening, evidence checks, clustering, and synthesis, while keeping final judgments, edge cases, and accountability firmly human-led.
- Explore attested access — reputation/attestation networks (e.g. using TrustGraph) could create permeable membranes where new entrants can join, but credibility is legible and review depth is matched to risk.
Handled well, these design choices collectively shape the “membrane” of the grant programs — determining how permeable it is to new entrants, how much scrutiny applications receive at different stages, and how operator capacity is protected without defaulting to closure or bureaucracy.