Gnome / Industry playbookCited briefing
A source may be current but use a different population or method.
Gnome / Industry playbook
Research
Coordinate source gathering, synthesis, and review for evidence-based briefs.
Planning guidance, not a claim of installed capabilities or customer results. Validate scope and feasibility before deployment.
Industry overview
Research teams need traceable evidence rather than fluent but unsupported summaries. A bounded agent workflow can separate source collection, comparison, and editorial review while keeping uncertain findings visible.
Conflicting evidence
Two estimates may disagree because their dates or definitions differ.
Evidence register
A broken source link makes a polished summary difficult to audit.
Use cases
01
Cited briefing
- Context
- A source may be current but use a different population or method.
- Human action
- Prepare a cited briefing from an approved source list.
- Intended outcome
- A draft whose material claims can be traced to source passages.
02
Conflicting evidence
- Context
- Two estimates may disagree because their dates or definitions differ.
- Human action
- Compare conflicting claims without silently choosing a winner.
- Intended outcome
- A conflict note preserving both methods for analyst judgement.
03
Evidence register
- Context
- A broken source link makes a polished summary difficult to audit.
- Human action
- Maintain an evidence table for a human reviewer.
- Intended outcome
- A reviewable table of claims, source versions, and unresolved gaps.
Agent, tool, and model capabilities
Separate tasks, preserve evidence
Plan separate retrieval and synthesis tasks, use permission-scoped document or browser tools, and select models against citation tests. Retrieved text is untrusted input, not authority to change the task or access policy.
Evidence contract
Keep source version, tool inputs, and execution status with each draft. Missing evidence remains visible.
A person owns the decision
A draft whose material claims can be traced to source passages.
Define acceptanceProposed capabilities must be tested against representative inputs. Unsupported or low-confidence results belong in a review queue, not an automatic decision.
Real-world scenario
Hypothetical, not a client story
Hypothetical: two reports disagree on a market estimate. A collection agent records dates and methods, a synthesis agent highlights the conflict, and an analyst decides whether either figure belongs in the brief.
- 01
Context
A source may be current but use a different population or method.
- 02
Review trigger
A draft or exception needs source verification before it can leave the workspace.
- 03
Human response
Prepare a cited briefing from an approved source list.
- 04
Intended result
A draft whose material claims can be traced to source passages.
How it works
Approved inputs
Receive a bounded task and permission-limited documents. Record source versions and reject access outside the agreed scope.
Agents and tools
Route retrieval, calculation, or drafting to selected models and scoped tools. Preserve failures, evidence, and approval boundaries in the run record.
Reviewed deliverable
An assigned person checks evidence, records the outcome, and follows the existing operational procedure. Rejected signals feed back into evaluation.
Business and operational value
Measure source coverage, citation accuracy, and time to human approval.
Evidence quality
Accepted observations / reviewed observations. Count missed cases separately against the manual reference.
Target: agree before pilotReview effort
Record minutes per reviewed item, including rework and escalations. Compare the same task with the manual baseline.
Target: agree before pilotSafe failure
Record whether stale inputs, denied access, and unavailable sources stop or visibly degrade the workflow.
Target: agree before pilotEstablish a manual baseline, agree acceptance thresholds with the operational owner, and compare review effort as well as accuracy. Any benefit must be measured in the pilot; no savings or ROI are promised here.
Deployment plan
Pilot one research question with a fixed source budget and approved corpus. Test missing citations and conflicting sources, assign an editorial approver, and stop runs that exceed access or cost limits.
- Scope: name the owner, permitted inputs, reviewers, and acceptance criteria.
- Prepare: assess local or approved hosted models, tool isolation, context limits, and a per-run cost budget.
- Validate: compare with human-labelled samples, exercise denied access and outages, and document rejected results.
- Decide: approve a limited rollout only after review; keep a rollback owner and re-evaluate when inputs change.
Existing tools and infrastructure
Review document formats, search access, browser restrictions, model hosting, and licensed-source terms. CCTV is not required; the prerequisite is authorized access to the research material.
| Check | Required evidence |
|---|---|
| Source access | Approved formats, source versions, licenses, and read permissions. |
| Tool boundaries | Sandbox, denied-command tests, credential isolation, and context limits. |
| Compute & recovery | Model endpoint, per-run budget, timeout behavior, and resumable evidence. |
A workflow is not a universal connector. Verify file formats, authentication, tool permissions, model context limits, and failure recovery in the actual environment before committing to a setup.
Integration planning
01
Approved source
02
Scoped processing
03
Human approval
04
Controlled export
Example: export a cited draft and evidence table into an approved document workspace. Validate source links and document permissions; publishing remains a separate editorial action.
These are example integration plans, not live connectors. Confirm schemas, least-privilege credentials, delivery acknowledgements, retry limits, and duplicate handling in a sandbox before enabling data exchange.
Security and privacy
Exclude restricted participant data unless specifically approved. Respect licensed material, limit prompt logging, and assess external model data handling before sending unpublished research.
Least privilege
Approve reviewer roles and test a denied-access case before launch.
Data lifecycle
Set retention, deletion ownership, and encryption for stored and transmitted evidence.
Accountable operations
Log access and decisions; rehearse incident escalation and rollback.
Before launch, approve purpose and lawful access, role-based permissions, encryption configuration, retention and deletion rules, audit logging, and incident ownership. Verify these controls in the chosen environment; this page does not claim compliance certification.
Hypothetical pilot example
HYPOTHETICAL PILOT
One bounded question. Evidence before expansion.
Hypothetical pilot: build a brief from a small approved corpus containing one deliberate contradiction. An analyst verifies each citation and records unsupported claims; completion alone is not acceptance.
A proposed evaluation exercise, not a deployed case study. There are no named clients, claimed results, or implied endorsements.
Frequently asked questions
Can a fluent answer replace citations?
No. Review source support for each material claim before accepting the brief.
Are paid sources automatically accessible?
No. Tool access must respect the team’s licenses and explicit permissions.
What if sources disagree?
Preserve the disagreement with dates and methods for the analyst to resolve.
Plan your next step
Bring one bounded task, approved inputs, tool permissions, and a named reviewer. Outline the team workflow in a planning brief before choosing models and execution resources.