Freeze the question
Replace the placeholders with one neutral buyer decision, audience, market, and constraint.
Free CSV · 18 fields · 12 prompt starters
AI Visibility Audit Evidence Log Template
Record what an AI answer actually showed before turning it into a score, claim, or client recommendation. The starter CSV keeps the prompt, engine, date, exact answer, citations, competitors, interpretation, owner, and next check in one inspectable register.
No email required. Opens in Excel, Google Sheets, Numbers, and most spreadsheet tools.
Boundary
This starter does not run prompts, does not automate monitoring, does not calculate a universal visibility score, and does not guarantee citations, rankings, traffic, leads, or revenue. It is a structured record for manual observations.
Use the register in five steps
Evidence first. Interpretation second.
A defensible audit preserves the conditions and answer before summarizing what the result may mean. The free CSV is deliberately simple enough to inspect, export, and challenge.
Record conditions
Capture the engine, model or mode, market, language, and calendar date.
Keep the answer
Store the exact response or a durable reference before scoring or rewriting it.
Separate observation
Record mentions, citations, competitors, and factual notes apart from interpretation.
Name the next check
Assign one supported action, an owner, and the date the observation should be repeated.
What the 18 columns capture
Enough context to inspect the claim.
The register keeps collection, observation, and decision fields distinct so a reviewer can see where evidence ends and interpretation begins.
Define the run
prompt_id
buyer_stage
prompt_text
engine
model_or_mode
market
language
run_date
Preserve what appeared
exact_answer_or_reference
brand_mentioned
citations_present
citation_urls
competitors_mentioned
factual_accuracy_notes
Make the handoff explicit
interpretation
next_action
owner
next_check_date
Twelve neutral starting points
Test buyer decisions—not vanity prompts.
Replace every bracketed placeholder. Keep the wording neutral enough that the prompt can reveal an absence, competitor, objection, or inaccurate claim.
Category choice
- Best [category] options for [audience]?
- Providers suitable for [use_case] in [market]?
- What should [audience] look for when choosing [category]?
Alternatives
- Compare [brand] and [competitor] for [use_case].
- What are the alternatives to [brand] for [audience]?
- Which [category] option is better for [constraint]?
Buying friction
- What are the limitations of [brand] for [use_case]?
- What risks should [audience] consider before buying [category]?
- Which providers support [requirement] without [constraint]?
Claim accuracy
- What does [brand] offer for [use_case]?
- Does [brand] support [requirement] in [market]?
- What sources verify [brand]'s claim about [claim]?
A single run is one dated observation, not a permanent ranking. Repeat the same controlled prompt set before claiming movement.
From one register to client delivery
Start free. Add the operating system when the work becomes commercial.
The CSV is useful for a self-audit or method evaluation. Signal OS adds the local workstation, action board, report builder, documentation, and edition-specific delivery assets.
Evidence log
$0The ungated CSV above for one controlled baseline.
- 18 evidence fields
- 12 neutral prompt starters
- No automated monitoring
Agency
$349Turn controlled observations into a repeatable, commercially licensed client service.
- Commercial client-use license
- Pricing and margin calculator
- 30-day roadmap and reporting workbook
Studio
$749Standardize the Agency workflow across multiple teammates.
- Internal multi-seat license
- Editable team training deck
- Roles, handoffs, and team QA manual
Founding prices end August 3, 2026 at 2:22 PM Pacific. Stripe verifies payment before releasing the purchased ZIP.