Inspect the work

A claim should come with something you can judge.

Six worked examples from one real procurement, read from the award record and from the GAO decision on its protest. Look at the source, distinguish evidence from inference, and ask whether the proposed next step is justified.

These are illustrative examples built from real government data and records: the Army Corps of Engineers' CASTLE-NET task order and the GAO decision on its protest. Every figure links to its source, as published on the date shown. They are not customer results, a live data feed, or product screenshots, and DRAI did not perform this work for any party to that procurement.

DRAI / Qualification briefIllustrative example · real government records

Who could bid at all

Illustrative example · CASTLE-NET · real government records

Request: TOR 47QFCA25R0014 · issued 2025-06-06
Vehicle: GSA Alliant 2 GWAC · FAR 16.505
Term: 1-year base and four 1-year options
Type: Hybrid cost-plus-award-fee and fixed-price

Evidence

The request went only to holders of the GSA Alliant 2 contract. The award record's number-of-offers field reports 6. GAO's decision says only that GSA “received multiple proposals.”

Inference

Vehicle access was the first gate, ahead of capability. For a firm without an Alliant 2 prime contract the only path was teaming. For a holder, six offers on $1.4 billion meant the barrier was the work, not the crowd.

Still unknown

Who the other four offerors were, and who teamed with whom. The record counts offers; it does not name them, and GAO names only the awardee and the protester.

Proposed next step

Confirm the vehicle path before anything else. Without a prime contract, open teaming conversations before spending on capture.

How this was found

Contract award dataPulled the award record for task order 47QFCA26F0004.Vehicle: GSA Alliant 2. Number of offers: 6. Type: hybrid cost-plus-award-fee and fixed-price.
Contract award dataPulled the parent contract, 47QTCK18D0036.An unrestricted GWAC, so only its holders could be asked.
Published decision (GAO)Read GAO's decision on the protest, B-424255.The request went to Alliant 2 holders under FAR 16.505; GAO says “multiple proposals”.
Company dataResolved the awardee's identity across award records.One vendor record, so its other orders on the vehicle could be counted.

Inspect the source

Offer count: the federal award record (FPDS) for task order 47QFCA26F0004, number-of-offers field, transaction last modified 2026-07-31. Eligibility: GAO, the request went “to holders of the GSA Alliant 2 governmentwide acquisition contract (GWAC), under the procedures of Federal Acquisition Regulation (FAR) section 16.505.”

GAO decision B-424255, B-424255.2 (19 May 2026) ↗

Human decision requiredFact → inference → decision

What makes it useful?

What evidence would change the investment decision?

Compare time to an approved qualification decision, evidence coverage, and the number of material unknowns resolved.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

DRAI / Customer brief + idea menuIllustrative example · real government records

What the customer said it would buy

Illustrative example · CASTLE-NET · real government records

Buys: U.S. Army Corps of Engineers, Office of the CIO/G-6
Awards: GSA, in support of the Army
Signal: Forecast as “formally RITS 2.0”; solicited as CASTLE-NET
Evaluation: 4 technical factors, descending · scenario in Factor 1

Evidence

GAO records four technical factors in descending order: technical approach and scenario response; management approach; key personnel and project staffing; corporate experience. Combined, they were “significantly more important than cost/price.”

Inference

The customer said in advance that it would buy reasoning over record: the scenario sat in the first factor and experience in the last. For the incumbent, years of relationship sat in the lowest-ranked factor; for a challenger, the scenario was the opening. That the new name signalled a modernization priority is a hypothesis.

Still unknown

Why the program was renamed, and what the scenario asked. Neither is in the public record.

Proposed next step

Take the idea menu to the customer before the final request, and listen for which idea they argue with. Rehearse the scenario during capture, not at proposal.

How this was found

Government entity & hierarchy dataLooked up the Corps of Engineers in the federal hierarchy, then the CIO office within it.The requiring office, distinct from the office that awards on its behalf.
Forecast dataFound the acquisition forecast entry for the recompete.Published 7 April 2025 as “formally RITS 2.0”, value band $500M to $999M.
Published decision (GAO)Extracted the evaluation structure from the protest decision.Four technical factors in descending order; combined, “significantly more important than cost/price”.
Public webSearched official USACE sources for the CIO office's stated priorities and leadership.Named priorities and appointments, each with a URL, to test the modernization hypothesis against.

Inspect the source

GAO lists the factors in that order and states that the technical factors combined were “significantly more important than cost/price.” The forecast entry describes the requirement as “formally RITS 2.0.”

GAO decision B-424255, B-424255.2 (19 May 2026) ↗

Customer validation neededFact → inference → decision

What makes it useful?

Which idea did the customer argue with, and what did that reveal?

Review source relevance, findings corrected or confirmed, and the usefulness of the resulting customer conversation.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

DRAI / Mission-to-solution traceIllustrative example · real government records

Where the award was decided

Illustrative example · CASTLE-NET · real government records

Factor 1: Technical and scenario · Excellent vs Good
Factor 2: Management approach · Acceptable vs Acceptable
Factor 3: Key personnel and staffing · Acceptable vs Acceptable
Factor 4: Corporate experience · Relevant vs Relevant

Evidence

On three of the four technical factors, the awardee (Accenture Federal Services) and the protester (SAIC) received the same rating. They differed on one: technical approach and scenario response, Excellent against Good.

Inference

The whole technical difference lived in Factor 1. A solution that is merely sound there ties; only a better answer to the scenario separates. That holds for a challenger and for an incumbent alike.

Still unknown

What in the scenario response earned Excellent. The public decision describes the protest, not the winning approach.

Proposed next step

Put the strongest solution people on the scenario, and compare at least three concepts against it before choosing one. Record why the others were set aside.

How this was found

Published decision (GAO)Extracted the ratings table from the protest decision.Both offerors rated on each of the four factors.
Published decision (GAO)Compared the two columns factor by factor.Three of four identical; Factor 1 differed: Excellent against Good.
Contract award dataWalked the predecessor order's modification history.Where the scope had grown over the prior term - the operating reality a scenario answer must fit.

Inspect the source

GAO's published ratings table: Excellent and Good on technical approach and scenario response; Acceptable for both on management approach and on key personnel; Relevant for both on corporate experience.

GAO decision B-424255, B-424255.2 (19 May 2026) ↗

Solution authority reviewFact → inference → decision

What makes it useful?

Which concept can the team deliver, prove, and test with the customer?

Assess whether each major concept links to a customer outcome, a requirement, a delivery mechanism, and credible proof.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

DRAI / Response structure + linked claimsIllustrative example · real government records

Sixty days from forecast to request

Illustrative example · CASTLE-NET · real government records

Forecast: Published 2025-04-07 · “formally RITS 2.0”
Forecast value: $500,000,000 to $999,000,000
Request: Issued 2025-06-06 · 60 days later
Evaluated prices: $1.40B and $1.43B

Evidence

The public forecast appeared on 7 April 2025 with an estimated value of $500 million to $999 million. The request followed on 6 June 2025, sixty days later. Both evaluated prices GAO records were above $1.4 billion.

Inference

A team that started at the forecast had two months. Shaping had to happen earlier, in market research, and a response there is how a firm becomes known before the clock starts. The forecast's value band was a floor, not a guide.

Still unknown

Whether a request for information or an industry day preceded the forecast. It is not in the sources reviewed.

Proposed next step

Treat the predecessor's end date, not the forecast, as the trigger. Answer market research when it appears, and hold every experience claim until its owner approves the evidence.

How this was found

Forecast dataFound the forecast entry and its publish date.7 April 2025, with a value band well below the prices later evaluated.
Opportunity, RFI, solicitation & amendment dataSearched for the request and any market research that preceded it.Request issued 6 June 2025; no earlier RFI found in the sources reviewed.
Published decision (GAO)Took the evaluated prices from the decision.$1,401,410,077 and $1,431,877,836 - both above the forecast band.
Contract award dataRead the predecessor's end date from its award record.The real trigger for the clock, ahead of the forecast.

Inspect the source

The forecast entry gives the publish date, the value range, and the description “formally RITS 2.0”, a recompete for the Corps of Engineers' Office of the CIO/G-6. The request date and the prices are from the GAO decision.

Public acquisition forecast, published 7 April 2025 ↗

Evidence owner reviewFact → inference → decision

What makes it useful?

Can each material claim be supported and released?

Compare practitioner time to an approved response and the share of relevant requests your team can complete.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

DRAI / C/L/M crosswalk + evaluator outlineIllustrative example · real government records

What a better Factor 1 was worth

Illustrative example · CASTLE-NET · real government records

Awardee price: $1,431,877,836
Protester price: $1,401,410,077
Difference: $30,467,759 · about 2.2% higher
Outcome: Protest denied · 2026-05-19

Evidence

The awardee's evaluated cost/price was $1,431,877,836; the protester's was $1,401,410,077. The award went to the higher price, by $30,467,759. GAO denied the protest, including the challenge to the best-value tradeoff.

Inference

With three factors tied, one step of rating on Factor 1 outweighed a $30 million price advantage. A page spent on Factor 1 earned more than a page spent anywhere else, and price did not recover the gap.

Still unknown

How the agency reasoned about the premium. The public decision summarizes the tradeoff; the source selection document is not public.

Proposed next step

Allocate pages by factor weight, heaviest to the scenario and its proof. Do not plan on price to recover a Factor 1 gap.

How this was found

Published decision (GAO)Took both evaluated cost/price figures from the decision.$1,431,877,836 (awardee) and $1,401,410,077 (protester).
Published decision (GAO)Read GAO's ruling on the best-value tradeoff challenge.Denied: the technical evaluation, the disparate-treatment claim and the tradeoff all upheld.
Published decision (GAO)Set the price difference against the one rating difference.$30,467,759 for one step on Factor 1; nothing else separated them.

Inspect the source

GAO's decision gives both evaluated cost/price figures and denies the challenges to the technical evaluation, the allegations of disparate treatment, and the best-value tradeoff.

GAO decision B-424255, B-424255.2 (19 May 2026) ↗

Proposal lead reviewFact → inference → decision

What makes it useful?

Does every proposed section earn its space against the actual instructions and evaluation criteria?

Review traceability and unresolved compliance gaps, then compare time to an approved outline and downstream rework.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

DRAI / Compliance matrix + section reviewIllustrative example · real government records

What the evaluator held to

Illustrative example · CASTLE-NET · real government records

Grounds: Technical evaluation · disparate treatment · best-value tradeoff
Outcome: All denied · 2026-05-19
Basis: Evaluation reasonable and consistent with the stated criteria
Gate: Not Acceptable on any factor makes the offer ineligible

Evidence

The protester challenged the technical evaluation, alleged disparate treatment, and challenged the best-value tradeoff. GAO denied all three, finding the agency's evaluation reasonable and consistent with the solicitation's stated criteria.

Inference

The evaluator applied the published factor language as written, and GAO held it to that language. The closest pre-submission proxy for the evaluation is a review scored against the same words - for a challenger and for an incumbent alike.

Still unknown

What either offeror's own color reviews found before submission. Internal reviews are not part of the public record.

Proposed next step

Run the final review as the evaluator will read it: score each section against the factor definitions, mark every requirement covered, partial or missing, and resolve every gap before submission.

How this was found

Published decision (GAO)Read the grounds the protester raised.The technical evaluation, alleged disparate treatment, and the best-value tradeoff.
Published decision (GAO)Read GAO's disposition of each ground.All denied; the evaluation was found reasonable and consistent with the stated criteria.
Published decision (GAO)Extracted the factor definitions and weighting as GAO restates them.Four technical factors, descending; combined, “significantly more important than cost/price”.

Inspect the source

GAO's decision denies the challenges to the technical evaluation, the disparate-treatment allegation, and the tradeoff, and restates the four factors and their weighting.

GAO decision B-424255, B-424255.2 (19 May 2026) ↗

Proposal lead approvalFact → inference → decision

What makes it useful?

Would an evaluator find every required element where the instructions say to put it?

Track requirements covered before the final review, gaps closed before submission, and the rework that reaches the last days of the schedule.

Is the source relevant and current?
Does the reasoning distinguish fact from inference?
Are important unknowns visible?
Is the next decision assigned to a human?

Make the judgment explicit.

Mark whether this illustrative brief is ready for the next step. The choice stays in this open page.

Why a protest decision

The evaluator wrote down what won. Almost nobody reads it.

When an award is protested, GAO publishes how the agency scored every factor and how it weighed price against them. It is the only public record of an evaluator's reasoning, and it is rarely read back into the next pursuit.

DRAI treats decisions like this one, and the award record beside it, as a data source: read, extracted, and carried into the next bid as evidence. The five cards above show what one decision yields.

See how each workflow uses it

Test the work

Now test the work against something you know.

Use a real customer or pursuit in an appropriate DRAI environment. Compare the result with your own knowledge before expanding the workflow.

Start a free trial