SaaS AI-Visibility Planning: A Practical Workstream Guide

Correction — 31 October 2026: the earlier version included unsupported research or outcome claims without supporting records. Those findings and dataset promises are withdrawn. This is practical guidance, not a verified original study or client result.

Define the business decision

Choose a product category and audience, then gather genuine sales and support questions. Separate branded questions from category discovery. Agree how enquiries and qualified demos will be counted before starting work.

Record a baseline

Save answers to a stable question panel with dates and assistant settings. Record linked sources, mentions without links and factual errors. Do not invent a zero baseline or treat a selected panel as total market share.

Improve priority pages

Clarify product fit, limitations, integrations and pricing where available. Publish comparisons that fairly describe alternatives and cite genuine evidence. Add only markup that matches visible facts, and keep modification dates tied to substantive edits.

Keep external identity accurate

Maintain genuine company and author profiles. Do not create Wikipedia or Wikidata entries solely for promotion or suggest that a knowledge panel can be purchased. Independent coverage must be genuinely earned and relevant.

Measure without a promised deadline

Repeat the recorded panel and keep actual referrals and qualified leads separate. No verified 47-citation outcome, 90-day success rate or payback claim is supplied here. A planning window organises work; it is not a predicted result.

Evidence needed for a real case study

Publish permission to describe the engagement, dated prompt logs, before-and-after source records, relevant page changes and the lead-attribution method. Until those records exist, this remains practical guidance rather than a client result.

Sources and further reading

Google guidance on AI features