Find your brand's strengths and gaps in AI answers

You need to know how AI presents your brand and where to focus next. A report can contain many metrics, but your first useful result is simpler: identify what is working, what deserves attention, and what the available evidence cannot yet answer.

Your outcome: a baseline for a defined audience and market, with a strength to preserve and a gap to investigate when the evidence supports them.

Choose your starting point

Your situation	Start here
You already have a relevant Brand Insights report	Open its Overview and check scope and date before step 2.
You have never measured this scope	Review the proposed Prompt Set and run the report in step 1.
You already have observed benchmark results	Use that defined benchmark scope for the relevant visibility question; do not rerun a report solely to follow this article.

You do not need to build Prompt Universe or a benchmark before every Brand Insights report. Use the measurement that answers your current question.

First useful task: take one relevant finding from the summary to a decision. You do not need to resolve every metric before using the report. If your main concern is that the brand is absent, continue with Find out why your brand is missing from AI answers after confirming the observation.

1. Make the baseline relevant to your business

Choose the product scope, audience, market, language, and AI systems relevant to the decision. "Our brand in AI" is too broad if you sell several products to different customers.

For a new Brand Insights report, review the system-generated Prompt Set. Keep questions that represent real customer needs. Adjust, add, or remove questions when a business requirement is missing or outside scope, using the controls available for that set. You do not need to rewrite a suitable selection.

Review usage and access information shown before execution. After the run, use the report's actual Prompt Set and evidence to understand what was measured.

For an existing report, check whether its scope and date still fit the task. A report from before a product launch can remain a useful historical baseline while being unsuitable for describing the new product today.

2. Read the summary before choosing a metric

Open Overview, read the summary, and review Priority Actions. Select the finding most relevant to your next business decision. If the report identifies several issues, begin with one that affects a product or audience you currently serve.

Use Intelligence to understand the finding:

What you want to know	What to inspect	Useful interpretation
Does AI introduce us when customers ask about the category?	Eligible discovery visibility evidence	Whether the brand appears without the question supplying its name
Does AI recommend us for the customer's need?	Recommendation evidence	Whether the answer actively supports choosing the brand
Does AI describe us accurately and usefully?	Portrayal, positioning, strengths, and limitations	Which claims or attributes need attention
Which sources appear in relevant answers?	Citation evidence	Which cited materials deserve closer inspection

The denominator and eligible question scope matter. A recommendation result and a brand-understanding result may describe different questions. Read them as different dimensions, not interchangeable grades.

3. Locate the strength or gap

Use the relevant breakdowns available in the report to see whether the finding concerns a particular system, need, or question group. Avoid extending a narrow result to the whole brand.

For example, strong category visibility can coexist with weak recommendations for a specific integration requirement. The useful next question is what happens around that requirement, rather than whether the brand is simply "good" or "bad" in AI.

If the report shows missing or insufficient evidence, write down the missing scope. Do not treat it as zero performance. If no material weakness is supported, preserving the current information and monitoring an important need can be the right decision.

4. Check the evidence that could change your decision

Open the relevant response in Evidence. Confirm the question, system, date, and the wording behind the finding. Read enough surrounding text to distinguish a passing mention from a recommendation and a factual statement from an opinion.

If the report says a capability is described incorrectly, verify the capability against current product information. If the intelligence does not match its own response evidence, preserve that example for support before acting on the finding.

5. Leave with a useful baseline

Illustrative example. A project-management brand is included in broad tool-selection answers. In the selected small-business integration questions, however, answers describe its accounting connection as export-only. The team verifies that an automated connection exists.

Its baseline note reads:

Scope: small-business project management, selected market and language, systems and dates in this report.
Strength: the brand appears in relevant category-selection answers.
Gap: the integration is described inaccurately in the identified question group.
Evidence: link to the finding and affected responses.
Next decision: determine whether to correct public integration information or investigate a source carrying the old claim.
Limit: this finding does not establish the same problem across every system or customer segment.

Use that short note in your team's existing planning tool if a handoff is needed. There is no need to copy the entire report.

You are done when you can describe the measured scope and explain the next justified action-or the specific evidence needed before choosing one. Running a report is the start; identifying a decision it supports is the useful finish.

Next: use Improve how AI describes and positions your brand for a portrayal issue, or Turn SpyderBot recommendations into an action plan to assign the next task.