From dense NOFO to a complete submission — in weeks, not months
A ~$16.4M inland river terminal modernization: an $11.25M federal request, an overmatched local share, and a complete rubric-aligned submission in weeks, not months — with AI fluency parsing the rubric, reconciling the budget, and running the math. Decision pending.
At a glance
A Midwest river-logistics operator came to Prosody Consulting with a $16.36 million terminal modernization project and a single goal: compete for a federal port-infrastructure grant, sponsored through a regional port district as the eligible public applicant. The catch was the calendar. A dense Notice of Funding Opportunity had just dropped, the scoring rubric had shifted from prior cycles, and the application window closed in weeks, not months.
The operator was no stranger to large capital projects, and what it wanted was a smarter way into this one. Prosody’s answer started with strategy, not paperwork: a federal funding route sized to the build, matched to a project that was genuinely ready to move. The route it chose — a small project at a small port — capped the federal ask at $11.25 million, which meant the remaining $5.11 million had to be committed by the operator and documented. That share works out to 31.22% of eligible project cost against a 20% statutory minimum: an overmatch, put on the table deliberately rather than trimmed to the floor.
Prosody delivered a complete, rubric-aligned application on time — a budget reconciled to the dollar, a quantified benefit case the program did not even require, and a narrative written to score. The difference-maker was not working longer hours. It was the same AI-fluency discipline that let a small team move with the research depth, drafting speed, and quality control of a much larger one.
The FY26 round has not been announced. What follows is an account of how the application was built, not a claim about how it will be scored.
“The work that used to eat the first two weeks of an engagement — reading the NOFO line by line, mapping every requirement to the rubric, wrangling the client’s data — now happens in days. That time goes back into strategy and writing.” — Prosody engagement lead
The challenge
Federal infrastructure grants are won and lost in the details. For this engagement, three pressures stacked at once.
The program had changed its rules between cycles. Scoring criteria were reweighted, and the cycle carried a fresh set of discretionary priorities the Secretary could apply at final selection — the kind of addition that never appears in last year’s template. An application built from that template would have argued to the wrong rubric and quietly lost points before a reviewer reached the merits.
The source material was scattered. Engineering estimates, equipment quotes, cost workbooks, and prior-cycle narratives lived across dozens of files in multiple folders and formats — the raw material for the application, but not yet organized into anything a reviewer could score.
And the timeline was unforgiving. Between the funding notice and the submission deadline, Prosody had to absorb a long, technical solicitation, reconcile the budget to the dollar, build the economic case, and produce a polished narrative in the sponsor’s voice — without sacrificing the rigor that distinguishes a fundable application from a hopeful one.
The methodology, applied
Prosody’s edge is not “using AI.” It is knowing exactly where AI accelerates expert judgment and where the expert still has to drive. The engagement ran on a few deliberate plays.
1. Turning a dense NOFO into a working rubric
Rather than read the solicitation once and hope to remember it, Prosody used AI to parse the full Notice of Funding Opportunity into a structured map: every eligibility requirement, every scoring criterion and its weight, every change from the prior cycle, and every place the new priorities had to surface in the narrative. The result was a living checklist that drove drafting decisions — so the application was built to the current rubric from the first sentence, not retrofitted to it at the end.
2. Making messy client data usable in hours
The operator’s data dump — cost workbooks, vendor quotes, engineering scopes — was read, cross-referenced, and reconciled with AI assistance, then verified by hand. Discrepancies between the headline project cost and the line-item rollup surfaced early, when they were cheap to fix, instead of during a final-week scramble. The budget that reached the application tied out: total project cost, federal request, and local match all reconciled to a single source of truth.
3. Drafting to the score, in the right voice
With the rubric mapped and the numbers locked, AI helped Prosody draft narrative sections that put the strongest evidence against the highest-value criteria — quantified benefits surfaced in the economic argument, project readiness framed against the program’s priorities, numbers carried through every section rather than asserted once and dropped. Prior-cycle materials informed tone and voice; the arguments themselves were built fresh for this project and this rubric.
4. Building the benefit case the program didn’t require
As a small project at a small port, the application was exempt from submitting a benefit-cost analysis. Prosody built the quantified analysis anyway, as a standalone Project Benefit Analysis exhibit.
The discipline was in what it refused to claim. Benefits were monetized only where USDOT’s own BCA Guidance supplies a parameter value — published federal figures for things like truck operating cost, driver time, and emissions damage, applied to a physical baseline of on-site vehicle movements the project eliminates. Everything else — reliability, redundancy, downstream effects on other operators — was presented qualitatively, per the Guidance, rather than assigned a number the method couldn’t support.
That line matters more than the total. An analysis nobody asked for is only worth including if a reviewer can trust it, and the fastest way to lose that trust is to monetize something the federal guidance doesn’t let you monetize. Modeling was AI-assisted, calculations were run programmatically and re-checked by hand, and every figure traces to a published parameter or a client source document.
5. A built-in verification pass — red-teaming the work
Every AI-assisted step was followed by a deliberate verification pass — a red-team review of the work. Names, places, and figures were checked against primary sources, claims were pressure-tested, and the math was re-run. This is a different discipline from how the work was done by hand: the source material is collected and logically structured once, and every number is traced back to its source rather than re-keyed and hoped over. The workflow treats AI as a fast, tireless first-drafter and researcher — and the red-team pass, never the AI, as the final word.
What was delivered
Prosody delivered a complete, on-time application that a small team could not have produced at the same depth on the same timeline by hand.
The submission led with a reconciled budget: $11.25 million federal request plus $5.11 million committed non-federal share, tying to a $16.36 million total project cost — the three numbers reconciling to a single workbook, with the local share documented by a letter of financial commitment rather than asserted. At 31.22% against a 20% minimum, the match was an overmatch on the record.
It carried a quantified benefit analysis the program did not require, disciplined about the line between what federal guidance lets you monetize and what it doesn’t. Every section was written to the current rubric — including the discretionary priorities introduced for this cycle, of which the application addressed two of three affirmatively and the third explicitly as not applicable, closing the gap a recycled template would have left open. And the whole package was produced in a compressed window that left more time for strategy and revision, not less.
Just as important, the engagement left Prosody with a repeatable system. The same plays — parse the solicitation, reconcile the data, draft to the score, verify everything — now travel to the next application, and the one after that.
Status
The application was submitted on time, in full, on May 29, 2026. The FY26 round has not been announced, and no award has been made. What Prosody controlled was the quality of the submission; the rest belongs to the program’s reviewers and its calendar. This page will be updated when the round is decided — in either direction.
Why it matters
For the organizations Prosody serves, the takeaway is straightforward. AI fluency, used well, does not replace grant expertise — it multiplies it. It means a deadline that once felt impossible becomes workable. It means the budget reconciles and the economic case holds up. It means more of the engagement is spent on the strategic judgment that actually wins awards, and less on the mechanical work of reading, sorting, and reconciling.
It does not mean a guaranteed award. No one can promise that, and a firm that does is telling you something about itself. What a disciplined process buys is that the application is decided on the project’s merits rather than on an avoidable error — a budget that doesn’t tie, a criterion argued to last year’s rubric, a number that can’t be traced. That is the part within anyone’s control, and it is the part Prosody is accountable for.
It also means the work costs less. A small senior team, multiplied by AI fluency, delivered this at a fraction of what a traditional engineering-led grant effort would have billed — with the fee on the table, not buried in the back end of the project.
That is the standard Prosody brings to every engagement: the rigor reviewers expect, delivered at a speed the calendar demands.
Same methodology, different proving ground each time. See it analyzed at scale across the federal funding landscape — prosodylabs.prosodyconsulting.com.
Project details in this case study have been anonymized to protect client confidentiality. Prosody Consulting develops competitive grant applications for infrastructure, logistics, and public-sector clients. To discuss your project, request a consultation.