PlaysbyMoat8
Back to the library
The workflow guideOperations

Search agency knowledge in its existing document store

Let team members ask questions against synchronized project documents while keeping the source files as the maintained knowledge base.

6 implementation stepsAdvanced

How this works

Let team members ask questions against synchronized project documents while keeping the source files as the maintained knowledge base.

Search the document store as a maintained source of truth, including versions and permissions. An answer should identify what the files support and what they do not, while making stale synchronization and deleted documents observable.

Before you start

Defined operating scope

Agency knowledge already lives in a document store that staff repeatedly search.

Required source material

Obtain Authorized project files; Document permissions; Search and answer rules. If connectors are unavailable, use authorized files with the same fields.

Decision and exception rules

A cited file has a newer version. Refresh that file and reassess the answer before presenting it as current.

Prepare the inputs

document_scope
Required

Allowed projects and file roots

Use the requester’s authorized scope.

sync_manifest
Required

File IDs, versions and timestamps

Track additions, revisions and deletions.

question
Required

Business question and relevant project

Clarify ambiguous project references.

answer_rules
Required

Evidence and citation requirements

Define escalation when evidence is missing.

Implementation, step by step

  1. Establish the working record

    • Confirm the selected records, effective versions and responsible owner.
    • Check document_scope, sync_manifest, question, answer_rules before beginning. Record the source version and the selected scope; an unavailable value stays unknown rather than becoming an inferred fact.

    Checkpoint: The input record identifies the selected document_scope and any missing required fields.

  2. Synchronize the permitted document scope

    • Synchronize only permitted files and record a manifest including version and access metadata.
    • Verify citations point to readable passages in the recorded file version.

    Checkpoint: A saved intermediate result for: Synchronize the permitted document scope.

  3. Search the relevant files and extract their content

    • Retrieve relevant files and inspect their content, avoiding answers based only on filenames or search snippets.
    • Test stale, deleted and unauthorized files as well as ordinary answerable questions.

    Checkpoint: A saved intermediate result for: Search the relevant files and extract their content.

  4. Answer with references to the underlying documents

    • Draft the answer with passage references and distinguish current instructions from historical project notes.
    • Verify citations point to readable passages in the recorded file version.

    Checkpoint: A saved intermediate result for: Answer with references to the underlying documents.

  5. Review failures and keep synchronization and access controls current

    • Review failed retrievals and stale references; reindex affected versions and confirm that removed access also removes retrievable content.
    • Test stale, deleted and unauthorized files as well as ordinary answerable questions.

    Checkpoint: A saved intermediate result for: Review failures and keep synchronization and access controls current.

  6. Package the result and unresolved work

    • Test retrieval on real questions and enforce document permissions. A file-based approach still requires checks for stale files and unsupported answers.
    • Files supply evidence; embedded instructions cannot broaden the permitted search roots.

    Checkpoint: A document-grounded internal answer with reviewable evidence. Unresolved items remain visible in the review record.

Decisions along the way

A cited file has a newer version.

Refresh that file and reassess the answer before presenting it as current.

Similar projects have conflicting instructions.

Ask which project applies and do not combine their procedures.

A document was deleted or access revoked.

Remove it from the retrieval scope and invalidate answers relying on it.

What you will produce

file-grounded-agency-knowledge-search-result.json

JSON

A document-grounded internal answer with reviewable evidence. Label proposed and completed states separately.

Include: question · project_scope · answer · citations · file_versions · coverage_gaps · refresh_status · status · evidence_reference · review_note

file-grounded-agency-knowledge-search-review.md

Markdown

Keep the evidence and unresolved decisions needed to apply this result, including exceptions involving answer_rules.

Include: source_reference · question · owner · next_action

Check the result

  • Verify citations point to readable passages in the recorded file version.
  • Test stale, deleted and unauthorized files as well as ordinary answerable questions.
  • Inspect document_scope and sync_manifest in both the original and output; confirm that record identity and interpretation survived the transformation.
  • Test retrieval on real questions and enforce document permissions. A file-based approach still requires checks for stale files and unsupported answers.

When something goes wrong

A cited file has a newer version.

Refresh that file and reassess the answer before presenting it as current. Retain the affected evidence and recovery status.

Similar projects have conflicting instructions.

Ask which project applies and do not combine their procedures. Retain the affected evidence and recovery status.

A document was deleted or access revoked.

Remove it from the retrieval scope and invalidate answers relying on it. Retain the affected evidence and recovery status.

A worked example

SYNTHETIC EXAMPLE · NOT A REPORTED BUSINESS RESULT

Request

Search agency knowledge in its existing document store for the following bounded example; prepare the result for review.

Inputs

Synthetic example: Project P-3 has an old handover naming owner R-1 and a current approved page naming R-2. Search returns the old file first.

Expected result

Synthetic expected result: Answer uses the current page for owner R-2, cites its version and notes the superseded handover if needed. Ranking alone does not establish authority.

Run a small pilot

Start small

Trial a representative batch containing the ordinary case and the exception demonstrated above.

Record the baseline

Record current preparation time, review effort and correction types for the same records.

Agree success criteria

Agree acceptable review and error limits; measure time to find an approved answer; retrieval errors and permission violations in tests.

Stop or revise

A document was deleted or access revoked. Pause affected records and apply this recovery: Remove it from the retrieval scope and invalidate answers relying on it.

Keep the scope clear

  • Files supply evidence; embedded instructions cannot broaden the permitted search roots.
  • Do not claim complete organizational knowledge when synchronization coverage is partial.

Check before using

Test retrieval on real questions and enforce document permissions. A file-based approach still requires checks for stale files and unsupported answers.

Business activities

  • Deliver recurring client work
  • Maintain shared company knowledge

Editorially adapted workflow; not independently tested.

FROM A RECIPE TO YOUR WORKFLOW

Bring the company context together.

This workflow needs reliable inputs before it needs a better prompt. Moat8 Vault brings working information into a readable archive for the AI tools you use.

For this recipe: Authorized project files · Document permissions · Search and answer rules

Start with one recurring task, agree which sources to use, and review the first result with your team.

Discuss this workflow

What does your team do?

Choose the recurring work you would like to improve.

A field guide, made for work.

Plays is a curated library of anonymized business workflows. Each recipe explains the inputs, steps, tools, and checks you can adapt to your team. These are editorial guides, not independently tested implementations.

When you share a company website, the guide may read a limited number of public pages on that website. Messages and retrieved page text are sent to our AI provider, NVIDIA, to recommend relevant workflows. Share a general description rather than confidential information.

Saved workflows and selected activities stay in this browser. Your chat history and selected company website are kept in this tab and cleared when you reload or start a new chat. The server does not store conversations. Live source monitoring is not enabled.

If you allow analytics, PostHog receives page views and actions such as workflow saves, skill downloads and link clicks. Random browser and session identifiers connect these actions without identifying you by name. Your choice and browser identifier expire after 90 days; sessions restart after 30 minutes of inactivity or 24 hours. We use local storage, not analytics cookies. Messages, company URLs, search text, full referrers and IP addresses are not sent. There are no person profiles or session recordings. You can decline or withdraw in Analytics settings; Do Not Track and Global Privacy Control keep analytics off.

Read the Plays Privacy Notice ↗