Skip to content

Capability 02 · Read the landscape

Competitive Intelligence

Build one evidence run for a defined competitor question, then reuse its fingerprinted records for a briefing and timeline.

The problem

Landscape updates are hard to compare when every briefing repeats different searches. A failed provider can also be mistaken for a trustworthy zero-result response, turning a coverage gap into a false market conclusion.

Objective

Collect trials, regulatory records, publications, and configured news evidence into an immutable run, expose source coverage and limitations, and render multiple business views from the same evidence fingerprint.

How it helps

  • Tracks explicit drug and company identities rather than relying on an ambiguous topic alone.
  • Distinguishes complete, partial, failed, not_configured, and not_applicable source states.
  • Binds report and timeline artifacts to the same run_records_sha256 so their evidence base can be compared.

How the plugin works

Input

Bounded market question

Tracked drug and company identities, aliases, therapeutic scope, source selection, and decision horizon.

Tools

Collect once, project many views

  • ci_status
  • ci_track
  • ci_refresh
  • ci_scan_trials
  • ci_trial_detail
  • ci_scan_regulatory
  • ci_scan_news
  • ci_scan_publications
  • ci_extract_events
  • ci_landscape
  • ci_report
  • ci_timeline
Expected output

Evidence run, briefing, and timeline

Immutable source records plus HTML, JSON, and CSV views bound to the same evidence fingerprint.

A three-step workflow

01 / BOUND

Define the watch question

Name the fictional or governed entities, aliases, therapeutic scope, sources, and decision horizon.

02 / COLLECT

Create one evidence run

Retrieve source records once and inspect coverage, record counts, limitations, and errors.

03 / PROJECT

Review consistent views

Generate a briefing and calendar-filtered timeline from the same immutable run fingerprint.

Sample input

The fictional YAML names one simulated drug and company, asks a bounded quarterly question, and makes the requested sources and output views explicit.

Download fictional input Inspect sample manifest

Interpreted sample output

The representative excerpt shows one simulated trial record under complete coverage and a not_configured web source. The trial observation can enter review; the missing web provider cannot be interpreted as evidence that no news exists. The briefing and timeline carry the same fictional run hash.

Download representative output JSON

Business value indicators

Indicator What a business reviewer learns
Coverage by entity and source Whether each requested evidence lane is trustworthy or inconclusive
Returned records versus total available Whether the displayed slice may omit relevant records
Run fingerprint reuse Whether briefing and timeline share the same evidence base
Undated or excluded events What the calendar view leaves out
High-impact findings confirmed Which observations have been checked against linked primary sources

Boundaries and human review

A complete source can legitimately return zero records; failed and not-configured sources cannot. Confirm high-impact findings against the linked primary record, assess query scope and retrieval time, and apply approved controls to stored evidence. This public-beta workflow supports qualified commercial and medical review; it does not prove that a market event did or did not occur outside the collected coverage.

Get started

Use the fictional run to practise interpreting coverage before findings, then consult the canonical pinned guide for provider setup and immutable artifact behavior.

Open the pinned Competitive Intelligence guide →

Utility links