Anonymised Research Case Study

The way a comparison was framed changed the answer.

In one AI Brand Intelligence Report™ measurement, the same company won every comparison anchored to its own strongest differentiator - and lost most of the open, unprompted "which is better" comparisons put to one engine.

Confidence: Incomplete 144 stored AI responses · 4 engines · anonymised B2B software company

This is an anonymised case study from a real AI Brand Intelligence Report™ measurement, run during SEOWEBI's initial testing programme. It presents findings from the original measurement period. It does not imply company endorsement, implementation of recommendations or measured business improvement.

The buying question being tested

Buyers ask AI engines to compare vendors in different ways - sometimes naming the attribute they care about, sometimes just asking which is better. This measurement tested both framings side by side: comparisons anchored to the company's own stated differentiator, and open comparisons that left the criteria for "better" up to the engine.

What SEOWEBI measured

36 prompts were run against each of four engines - ChatGPT, Gemini, Perplexity and Claude - producing 144 stored responses. Two comparison subsets matter here: four prompts anchoring the comparison to the company's own differentiator, run against all four engines (16 responses), and four open "which is better" prompts against the same named competitors, also run against all four engines (16 responses). Every response was stored with its prompt, engine and timestamp.

Differentiator-anchored comparisons — 16 responses

Open "which is better" comparisons — 16 responses

Each square is one stored comparison response. Every differentiator-anchored response favoured the company (top panel). In the open comparisons (bottom panel), one engine named a competitor the better choice in three of its four pairings - the three orange squares; the remaining nine open-comparison responses, from the other three engines, framed the choice as depending on use case.

16 of 16

anchored responses favoured the company

3 of 4

one engine's open comparisons favoured a competitor

engines tested

The central finding

When the comparison was framed around the company's own differentiator, all four engines favoured the company in every one of sixteen responses. When the same four competitor pairings were put as an open "which is better" question, one engine named a competitor the better default choice in three of its four responses.

Two supporting facts sit behind that finding. First, the anchored result was unanimous across all four engines tested, not just the one that later diverged on the open framing - the differentiator-based framing produced a consistent, uncontested outcome. Second, the other three engines did not follow the same pattern on the open comparisons: they described the choice between vendors as depending on the buyer's use case, rather than declaring a winner - so the divergence sits with one engine's behaviour on the open framing, not across all four.

Why it matters

Observation: the same underlying comparison produced a different outcome depending on how the question was framed - anchored to the company's own attribute, or left open for the engine to decide the criteria.

Interpretation: this suggests that published content emphasising the company's differentiator plausibly shapes anchored answers, while nothing currently supplies that framing when a buyer asks an open question. It does not establish an overall AI "win rate" against these competitors - the anchored and open results measure different questions, and combining them into a single score would overstate what either subset shows on its own.

The priority the report recommended

The report's 90-day roadmap listed resolving the open-comparison gap as a medium-priority action for days 30-60, framed in the report's own words: on the four comparisons framed as an open "which is better," one engine names a competitor the better default choice for most teams in three of four cases; the same pairings framed around the company's own differentiator favour the company unanimously. Published content that leads with the attribute-based framing, rather than waiting for the buyer to supply it, was identified as the most direct route to closing that gap. This reflects what the report recommended at the time of delivery. SEOWEBI has no record of any such content change, or a remeasurement, having taken place since.

Measurement limitations and what remains unknown

This report's overall confidence label is Incomplete: a small number of sub-signals, worth a minority of the total 100-point score, could not be verified from available data and were scored conservatively as zero rather than estimated. Because of this gap, the report's score range crosses a scoring-band boundary, so this measurement does not settle which band the company sits in overall - a plain-language range is the accurate way to describe it, not a single score. Both comparison layers described above were fully and completely measured, with no gaps: all 16 anchored and all 16 open-comparison prompt-engine pairs returned a stored response. What remains unknown is whether this pattern has since changed: this case study describes AI-generated answers recorded during a single measurement window in 2026, not current or ongoing AI behaviour.

Where this fits

This finding is one expression of the pattern SEOWEBI calls Incumbent Squeeze - a brand crowded out of category-defining answers despite accurate description once it is named. All three case studies in this collection identify the same underlying pattern from a different angle.

This is what an AI Brand Intelligence Report™ is built to surface.

The same evidence standard, confidence labelling and traceability shown in this case study applies to every AI Brand Intelligence Report™ - run against your own company and named competitors.

Run Your AI Brand Intelligence Report™ See What the Report Includes →

More anonymised research: Pricing discrepancies · Description vs. recommendation · All case studies