That’s where Karisha Stanley‘s story starts. Karisha is the business development analyst at Contract Management Services (CMS), the revenue-generating arm of the Bremerton Housing Authority in Washington State. She’s also capture lead, proposal manager and writer, all in one. In our recent customer webinar, Shredding the Competition, she and quality assurance analyst Michele Raymond walked through how their team went from proposal chaos to what Karisha calls “organized chaos.”
The most surprising lesson didn’t come from a win. It came from a loss, and from what the winning proposal revealed about theirs.
The perfect storm
CMS first used VisibleThread for document comparison. Its compliance team audits properties on behalf of HUD, and policy documents can run past 100 pages. Comparing this year’s version against last year’s showed exactly what changed before reviewers arrived on site.
Then Karisha moved into business development. It was summer, she had PTO booked, and an RFQ dropped with a short turnaround. Writers, quality assurance and marketing had to respond without her.
Under pressure, the team did what most teams do: they fell back on old habits.
- Marketing reused old content, campaigns and visuals because that’s what was on hand.
- Writers didn’t feel confident. They had no structure telling them what the evaluator wanted or what a good answer looked like.
- Michele was running analyses and building a compliance matrix to plug gaps, but quality reviews had little follow-up because there was no time.
The audience recognized the pattern. In a live poll, 50% said they pull old content from past proposals when a tight RFP lands. A third said their team hits “good sports mode” on almost every pursuit, and more than half hit it a few times a year.
CMS submitted a complete response. But when Karisha came back, the verdict from the team was clear: they never wanted to do it again. As one attendee put it in the chat, they were celebrating getting it done rather than getting it right.
That was a problem, because doing it again is the whole point of the role. The revenue CMS wins funds new housing units, rental assistance and youth programs in its community.
Hitting reset: finding what the RFP doesn’t say
When the next opportunity came, Karisha set out to spare the team the pain of the last round. The first step was uncovering hot buttons: the concerns a customer cares about but never states outright.
If you helped shape an RFP, you probably know these already. If you found it on a procurement site, you have to dig. Karisha used VisibleThread’s AI as a research assistant, asking it questions grounded in the RFP: what might I be missing, and what would an evaluator prioritize? She then shaped and challenged the output with her own knowledge.
Three hot buttons surfaced for a HUD rent adjustment processing opportunity:
- Processing accuracy and regulatory compliance. Errors carry real financial and legal consequences. Evaluators would look closely at quality control architecture, so “we use a checklist” wasn’t enough. Writers had to explain how the checklist improves accuracy and reduces risk.
- Transition risk and continuity of operations. Processing follows fixed annual windows. A missed deadline can mean a resident pays the wrong rent. Evaluators would want transition plans and knowledge transfer: what happens if a dedicated processor is suddenly out?
- Staffing stability and key personnel. Evaluators would ask whether the named staff are real, committed and experienced in HUD Multifamily work, and whether the retention strategy is credible.
Each hot button became a win theme that Karisha gave to writers, so they knew what messaging to include.
How did Karisha decide which hot buttons mattered and which were noise? Risk. “In BD, it always comes down to risk,” she said. Focus on whatever could stop you from delivering for the customer. And when an attendee asked about bias, she was clear that an AI-flagged hot button is a lead, not a decision. Validate it against board minutes, past RFPs and the customer’s own website.
From shred to compliance matrix: guardrails for writers
Next came the shred. Karisha pulled every must, shall and should statement out of the RFP. These are the requirements that can disqualify you on their own. If proposals must arrive by 4:00 PM and yours lands at 4:15, you’re out. Federal rules work the same way: FAR 15.208 governs the submission, modification and withdrawal of proposals, and late submissions are generally not considered. GAO bid protest decisions regularly turn on missed requirements like this.
The shred does double duty. This is where RFP automation pays off first, because pulling every must, shall and should statement by hand is slow and easy to get wrong. Before any writing starts, it tells leadership what the customer actually requires, which makes the go/no-go decision an informed one.
If the answer is go, the shred becomes the compliance matrix. Each requirement gets an owner, a status, a reviewer and a win theme. Karisha also used VisibleThread prompting to add a “good response” note to each row. For example, a requirement to review rent adjustment requests for completeness came with guidance to describe how personnel, supervision, equipment and resources support that review.
This was the structure the first team had been missing. Writers knew their section, the requirement, the hidden concern behind it and what a high-scoring answer looked like.
Gap analysis and proof points, while writing
CMS didn’t wait for red team to check its work. As drafts came in, the team compared them against the evaluator expectations written into the RFP. That mattered because, like half the audience, they had been reusing language from past proposals, including ones they lost.
They ran the gap analysis several ways:
- By evaluation section. The cost proposal, weighted at 30%, came back “Moderate,” with a recommendation to add a worked example showing the estimated dollar range.
- By requirement. Their own words were quoted back and rated “Weak.” Extension pricing was identical to the base period, with no explanation.
- Against an outstanding response. Their 25-year record administering HUD programs across multiple states was rated “Outstanding. No gap.”
With time tight, they fixed the weakest areas first. Karisha stressed that the AI’s recommended fix is only a suggestion: the human decides whether to act on it.
The second check was for unsupported claims. “Proof points, proof points” became Karisha’s refrain. One draft said a standard contingency protocol would activate a qualified backup processor if someone left. The review flagged it: no named backup, no documented procedure or activation time, and a promise of no onboarding delay with nothing to back it up.
Having a neutral third party flag the claim helped. It wasn’t just Karisha or Michele asking for evidence anymore.
The audience agreed on timing. In the live poll, most said they check against evaluation criteria before writing starts or while drafting. Only 18.5% wait until the draft is done.
What the winning proposal revealed about ours
The new process worked. The team collaborated well, writers felt supported, and the compliance matrix was cited again and again in the after-action review. CMS still lost. It came down to cost, which is the customer’s budget and out of their control.
But Karisha didn’t stop there. She filed a public records request for the evaluation scores, the winning proposal and every other submission. Then she used VisibleThread to compare CMS’s response with the winner’s, criterion by criterion.
Some results were reassuring. On HUD systems proficiency, both proposals scored 5 out of 5. Both showed deep familiarity with TRACS, iREMS and the other systems the work relies on.
Others were humbling. On specific processing timelines, CMS scored 4 and the winner scored 5. CMS referred to “established timelines” and “HUD-established deadlines” without giving day counts. The competitor spelled them out: initial review within 5 business days of receipt, deficiency follow-up every 10 business days, calculation validation within 7 business days.
Same work, same pool of providers. The difference was how precisely each team described it.
That’s the lesson Karisha took away. If a bid comes down to price, let price be the only reason you lose, not a vague description of work you do every day. The competitor’s proposal also became a persuasive internal tool. It’s far easier to get a team to change habits, like turning dense text into visuals, when you can show how the winner did it.
How to get competitor proposals
Asked how to do this, Karisha shared her process. Wait until the award is announced. Then go to the issuing agency’s website and find its public records request portal or procurement team. Ask for the evaluation scores, the winning proposal and the other submissions. The agency will send what the law allows and explain anything it withholds. Rules vary by jurisdiction, so check what applies where you bid.
Five lessons to take into your next proposal
- Shred every solicitation the day it arrives. It’s the fastest route to an informed go/no-go decision.
- Turn the shred into a compliance matrix with guidance. Give each writer a requirement, a win theme and a picture of a strong answer.
- Look for the hot buttons the RFP doesn’t state. Prioritize by risk, and validate before you build win themes on them.
- Run gap analysis while you write, not after. Check every claim for a proof point.
- Learn from every loss. Request the winning proposal and scores, and compare criterion by criterion.
Asked to pick just one, both Karisha and Michele chose gap analysis. A gap can disqualify you before an evaluator reads your best material. The audience leaned the same way, with learning from competitor responses and gap analysis while writing as the top picks in the final poll.
Key takeaways: RFP automation in practice
- RFP automation covers shredding, compliance matrices and gap analysis, not just document drafting.
- CMS scored 5 out of 5 on HUD systems proficiency but 4 on processing timelines, where the winner gave day counts.
- In the live poll, 50% of attendees reuse old content when a tight RFP lands, and only 18.5% wait until the draft is done to check against evaluation criteria.
- Humans decide on every AI recommendation. Validate hot buttons and fixes before acting.
Advice for anyone new to proposals
Start by getting comfortable with shredding and building a compliance matrix. They’re your first guardrails. Then learn to support your technical writers. They know the work, but describing it is a different skill. Give them starting resources and a template that shows what you’re looking for. That’s how you build champions who want to do it again.
The proof is in the after-action review. After the first RFQ, nobody wanted to be called on again. After the second pursuit, even in a loss, the team’s response was: sign me up.
Want to see it in action? Watch the full Shredding the Competition webinar on demand, or book a demo to see how VisibleThread supports shredding, compliance matrices, gap analysis and competitor review.
