Codex
Claude
GeminiWhat you do
- State the business goal, audience, product proximity, markets, and topics to avoid.
- Choose approved keyword and competitor data sources.
- Review the opportunity map and select content priorities.
SEO and content teams choosing what to write next / seo
SEO Keyword Research
A prioritized opportunity map that separates near-term article ideas, supporting clusters, evidence, and topics to avoid.
From request to resultA complete plan you can follow.
Visible planThe problem
Keyword research produces broad high-volume lists when it is not anchored to the business, audience, product proximity, and existing authority.
AI can connect keyword data, SERP patterns, competitor evidence, and project context while the operator owns positioning and priority.
What the agent does
- Translate the business brief into explicit opportunity criteria before collecting data.
- Use approved keyword, SERP, and competitor sources and preserve provenance.
- Cluster by intent and business fit, score with transparent tradeoffs, and propose concrete article angles.
When this workflow helps
A reusable method, adapted to the request.
- A project needs SEO opportunities from seed terms, competitors, or an agent-defined research angle.
Before it starts
Connection required
- A business goal, audience, market, and product-proximity boundary
- Topics to include and avoid
- Connected keyword and SERP data sources or approved exports
What proves it worked
Evidence, not a success claim.
- Source-linked keyword set
- Intent and business-fit clusters
- Transparent priority scoring
- Concrete article angles and avoidance list
Workflow path
The reusable stages of the work.
A real run expands these stages around the request, context, selected tools, approvals, and dependencies.
- 01Passes to next stage
Orient From History
Read recent research, learnings, and the current project goal; if the user did not provide seed terms or competitors, ask a small number of…
- 02Passes to next stage
Collect Evidence
Execute only the provider calls the run plan grants and the agent still needs; preserve returned response-file paths, use schema.get with schema_ref only…
- 03Passes to next stage
Store Opportunities
Persist a useful list of keyword opportunities, not just one or two terms, with clusters, metrics, source refs, and a polished brief/CSV when the run…
- 04Verified outcome
Capture Learning
Record reusable lessons or rejected patterns for future runs.
Safe stopping and recovery
Useful even when the whole path cannot run.
- Stop after the prioritized opportunity map; article production is a separate approved job.
- If a provider is unavailable, use approved exports or stop with the exact missing dataset.
- Do not fill the report with broad terms that fail the product-proximity or authority test.
Documented follow-on work
The next workflow is conditional, not hidden.
Specialists inside this workflow
Clear roles for each part of the job.
Connected work