# Try a sourced VC research brief with fictional company data.

Reviewed: 2026-10-25

Canonical: https://clankercloud.ai/templates/vc-research-brief

Use this fictional Cedar Orbit packet to test a VC research brief in Workspace. Ask for a saved memo and evidence table that preserve reported metrics, source status, missing facts, and a later correction. The expected results are authored checks, not an executed Agent trial or verified company data.

## What you are working toward

Your own assessment of a saved brief’s source fidelity, calculations, missing-data behavior, and revision consistency.

## Start with

- A new Project for the fictional Cedar Orbit packet

- The initial metrics and source notes; withhold the correction initially

- A reviewer who checks the saved files against the authored reference

## Use the first two files for the initial task

Add reported-metrics.csv and source-notes.md to the Project. Ask for a saved brief and supporting table. Keep the dated correction for a second pass. The data is fictional, company-reported, and not independently verified.

The initial packet contains two quarters of revenue and active-business-customer counts. The notes qualify the figures and identify information that is missing. An absent metric should remain unknown.

## Authored reviewer reference

These checks describe the supplied packet. Record pass, revise, or not observed with the actual output file and version; do not count a confident explanation without a saved file as a successful artifact.

| Check | Expected reading |
| --- | --- |
| Initial quarterly revenue | Q2: $120,000. Q3: $150,000. Reported increase: $30,000, or 25%. |
| Revenue qualification | Q3 includes $25,000 of one-time setup revenue. Do not describe all $150,000 as recurring or assert an ARR. |
| Customer counts | 24 at Q2 end and 30 at Q3 end. Churn, customer cohorts, and retention are not provided. |
| Plans and verification | The European launch is proposed; approval and owner are unknown. Company figures are not independently verified. |
| Second-pass correction | Q3 revenue becomes $142,000: increase $22,000, about 18.3%. One-time setup revenue stays $25,000. |

## Apply the dated correction to every dependent claim

Supply dated-correction.md after reviewing the first result. It removes an $8,000 duplicate from the Q3 revenue export. Customer counts, setup revenue, and the proposal’s approval state are unchanged.

Ask for new saved versions of the memo and evidence table, plus a short change note. Check that no old 25% headline remains. A corrected table with an unchanged summary is an incomplete revision.

## Keep the questions and the reviewed version together

Save questions about the missing definitions, customer cohorts, cash data, and launch approval. Attach the reviewed file versions and next action to a Planner task if you are evaluating a handoff. Creating or assigning the task does not start another run.

If the method is useful, turn its source-checking rules into a reusable Agent. Keep the fictional company facts as test inputs rather than general reference knowledge. Try another packet before reusing the role for consequential work.

## Sample files

Fictional inputs. Download and add to your Project.

- [reported-metrics.csv](https://clankercloud.ai/templates/vc-research-brief-files/reported-metrics.csv): Fictional company-reported revenue and end-of-quarter customer counts.

- [source-notes.md](https://clankercloud.ai/templates/vc-research-brief-files/source-notes.md): Fictional qualifications, a proposal, and facts that are absent.

- [dated-correction.md](https://clankercloud.ai/templates/vc-research-brief-files/dated-correction.md): Second-pass correction to the fictional Q3 export.

## Try this prompt

Copy into a Workspace Session. Opening Workspace does not send it automatically.

```text
Use reported-metrics.csv and source-notes.md to save a research brief and supporting evidence table for our fictional Cedar Orbit review. Separate company-reported facts, qualifications, calculations, and missing information. Compare the two quarters without inventing recurring-revenue or retention claims. Include source references and follow-up questions. Save and inspect both files. Do not use a later correction until supplied, recommend an investment, contact anyone, or share files.
```

## Illustrative result

Fictional sample; your result depends on your inputs and needs review.

```text
Authored reference, not an Agent result: initial revenue rises from $120,000 to $150,000 (+25%). Q3 includes $25,000 one-time setup revenue; ARR and retention remain unknown. After the correction, Q3 is $142,000 (+about 18.3%). The launch remains proposed and the figures remain company-reported.
```

## Before using the result

- Initial comparison is 25%; corrected comparison is about 18.3%.

- One-time revenue is visible; no ARR or retention is invented.

- The proposed launch does not become an approved commitment.

- Company-reported facts remain distinct from independently checked facts.

- Both saved files and the change note agree after the correction.

## Common questions

### Is this a real investment case or measured Agent result?

No. Cedar Orbit and the source packet are fictional. The expected results are authored checks. Run the task yourself and record the saved files and behavior you observe.

### Can this test a reusable diligence Agent?

It can test a bounded research and source-checking role. Use the same criteria, withhold the correction initially, and inspect the actual saved results. It does not establish full diligence coverage or investment judgment.

## Sources

- [VC research workflow](https://clankercloud.ai/learn/ai-workspace-for-vcs)

- [Agent evaluation](https://clankercloud.ai/learn/test-ai-agent-before-reuse)

- [Saved Office deliverables](https://clankercloud.ai/learn/create-office-deliverables-with-ai)

## Related

- [An AI workspace for VC research and portfolio work.](https://clankercloud.ai/learn/ai-workspace-for-vcs)

- [An AI workspace for startups and founders.](https://clankercloud.ai/learn/ai-workspace-for-startups)

- [Give the Agent a test with an answer you can check.](https://clankercloud.ai/templates/agent-acceptance-test)

- [Keep the task, progress, and deliverable together in Planner.](https://clankercloud.ai/learn/ai-planner-tasks-and-deliverables)

- [Build AI Agents for the work your business repeats.](https://clankercloud.ai/learn/ai-agent-factory)

[Open Workspace](https://workspace.clankercloud.ai)

[Current pricing and usage](https://clankercloud.ai/pricing)