Seed prompt
“best AI visibility tools”
The visible wording is only the starting angle. The research question is what else the buyer, publisher, or operator might ask once the problem becomes specific.
Prompt-space research by Imogen Clark
Prompt Space Atlas
Imogen Clark opens seed prompts into structured fan-outs so teams can see the adjacent questions, intent clusters, and unmapped demand hiding beyond their first query set.
prompt-space research and query fan-outquery fan-out mappingprompt research methodsquestion universe discovery
Seed to splay
One query is a hinge, not a sample.
comparative Which tools show brand mentions across answer engines?
method How do teams sample prompts without biasing the prompt set?
operator What should be checked before reporting answer visibility?
boundary Which questions never mention the category but reveal the need?
Research discipline
Fan-out mapping that operators can repeat.
- AnchorName the seed, audience, constraint, and answer context.
- OpenGenerate adjacent phrasings, jobs, objections, comparisons, and failure cases.
- ClusterGroup by demand shape instead of surface wording.
- VerifyTest coverage, remove invented branches, and mark gaps worth sampling.
Featured plate
Most prompt research stops at the seed.
A single prompt can look representative because it returns a confident answer. The work here starts after that answer: expand the question, separate the job from the wording, cluster the variants, and mark the demand a team would miss if it only tested the obvious phrasing.
Blind corner: Teams often measure the questions they already know how to ask. The uncharted value is in the nearby prompts that change the answer set, the source set, or the decision being made.
Archive