Why letters of support matter — and when they don't
Letters of support (LOS) are optional at most agencies but carry real signaling weight when strong. A letter from an actual paying customer or program office end-user is one of the few things that can move a selectable proposal into selected. A stack of weak letters — generic industry consultants, academic collaborators with no funding tie — often hurts rather than helps. Reviewers read letters and they penalize padding.
Letter of Support Collection Process
- Step 1 — Identify the right signers: Letters from future customers score higher than letters from research partners. The best letters come from the program office that will eventually buy.
- Step 2 — Write the letter for them: Send a draft. Letters that arrive by email with "feel free to edit" produce stronger results than asking partners to write from scratch.
- Step 3 — Include specific commitments: "We will provide access to our production dataset during Phase I" scores better than "we think this is important research."
- Step 4 — Match the letter to the commercialization plan: Every letter should support a named customer or use-case in your commercialization narrative. Orphan letters add no score.
The hierarchy of signaling strength

| Letter type | Signal strength | Why |
|---|---|---|
| Paying commercial customer (LOI) | Very strong | Committed money. Nothing tests commercial viability better. |
| Program office with funding identified | Very strong | Named transition path with dollars behind it. |
| Program office without funding | Strong | Government customer interest. Suggests Phase III pull-through. |
| Large prime with teaming agreement | Moderate | Signals ability to execute at scale. Depends on specificity. |
| Prime without teaming agreement | Weak-moderate | Depends on the prime and the specificity of the interest. |
| Academic collaborator | Weak | Useful only if collaborator brings distinctive data or method. |
| Industry association / trade group | Weak | Rarely adds meaningful signal. |
| Vendor or supplier | Very weak | Can be reversed — "vendor lock-in" reads badly. |
What a strong letter contains
A strong letter has four elements: who the writer is and why they matter, what specific problem they have, why this firm and this approach solves the problem, and what they would do if the Phase I / Phase II is awarded. The writer should know enough about your technical approach to describe it in one sentence. If the letter could be cut-and-pasted to another firm's proposal with only the name changed, it is generic and reads as such.
Concrete example of strong language: "As the maintenance operations manager for the 25th Infantry Division, I am responsible for readiness on 120 rotary-wing aircraft. Our current predictive maintenance approach produces false positives at a rate that consumes over 400 hours per month of mechanic time. Based on the briefing Mr. Peng provided in March 2026 and a review of his federal health agency production ML work, I believe his approach to anomaly detection with calibrated uncertainty can materially reduce our false-positive rate. If this Phase I is awarded and feasibility is demonstrated, we would evaluate the Phase II prototype on our fleet data and engage the G-4 staff on funding a Phase III transition."
That letter tells the reviewer: who has the problem, how big the problem is, why this firm is credible, and what the transition looks like. A letter that says "we are excited to collaborate with Precision Federal on innovative AI" tells them nothing.
How to request a letter
The ask: a short email to a named contact who you have already had a substantive technical conversation with. Draft the letter for them in their voice, send it as an editable Word document, ask for feedback or signature on letterhead. Most people will sign a well-drafted letter within a week. Nobody has time to write one from scratch.
Script for the ask: "Dr. X, thank you for the conversation on [specific date] about [specific problem]. We are submitting a Phase I proposal on [topic] and a letter of support from you would strengthen the commercialization case. I have drafted language reflecting our conversation — would you be willing to sign on your letterhead, with any edits you want to make? Happy to revise."
What makes reviewers dismiss a letter
- Generic boilerplate. No specific problem, no specific approach, no specific commitment.
- Form letter from multiple signers. Same text signed by 5 different people reads as a mass mailing.
- Conflict of interest. Letter from a paid consultant on the project reads as self-dealing.
- Unknown signatory. A title that cannot be verified or a role that does not align with decision authority.
- Overreach. Commitments the signer cannot credibly make (e.g., a program analyst committing Phase III dollars).
- Too many letters. Stacks of 10+ weak letters read as insecurity.
How many letters to include
Phase I: 2-4 strong letters. Phase II: 3-6 strong letters. Quality over quantity — if you only have 1 strong letter and 5 weak ones, submit the 1 strong letter and skip the rest. Reviewers respect discipline.
The federal letter — specific patterns
Program office letters have specific conventions. They usually come on agency letterhead, reference the specific topic number, and cite the writer's role and decision authority. They do not commit future funds — agencies cannot legally pre-commit funding — but they can express a specific transition interest, identify a program of record that would benefit, and name follow-on activities that would occur if the capability is demonstrated.
Agencies vary on whether program office letters are welcomed, allowed, or discouraged. DoD generally welcomes them. NIH and NSF are more neutral. Check the solicitation and, if uncertain, ask the TPOC.
Timing the request
Letters take 1-4 weeks to secure. Start the ask 30 days before the proposal is due. For a DoD cycle where proposals are due 60 days after solicitation release, that means initiating conversations in the first two weeks of the cycle. Firms that wait until the final two weeks often submit weaker letters because the signer did not have time to review the technical approach.
Storing letters for reuse
Non-topic-specific letters (about the firm's capability generally) can be reused across proposals. Topic-specific letters cannot. Keep a repository of signed letters with metadata (signer, date, scope) and request renewals for letters older than 12 months.
Frequently asked questions
2-4 strong letters for Phase I, 3-6 for Phase II. Quality over quantity.
No, but they can express transition interest and identify a program of record.
Only if the academic brings distinctive data, method, or capability. Generic academic letters are weak.
Yes, always. Letters without letterhead and signature are discounted.
Standard practice. Draft in their voice, send as editable document, accept their edits.
30+ days before the proposal is due. Letters take 1-4 weeks to secure.