Platform · Impact · Case study

US B2B SaaS vendor: 128 → 175 answers naming it, out of 1,000.

A vendor of IT service management software, selling in the United States. Two identical measurements, on August 9, 2026 and September 22, 2026. In between: an action plan, technical fixes and 3 published assets.

Anonymized customer44 days13 targeted questions3 assets published

Starting point · August 9, 2026

128 answers out of 1,000 named the brand

Named 128Source only 110Absent 762

16 of the 50 buying questions never named the brand. Most-cited sources in this market: youtube.com (390), reddit.com (373), out of 12,985 citations.

What we did

44 days, five steps, the same protocol at the start and at the end.

Baseline measurementAugust 9, 2026. 50 buying questions, 1,000 assistant answers, 250 Google searches.
Measurement engine
Action plan13 questions targeted first, where the brand was most absent.
Experts
Technical fixesTechnical score raised from 73 to 85 out of 100.
Audit + web team
Production3 assets published: a pillar asset, a compliance asset, a video.
ZivRank Engine
Re-measurementSeptember 22, 2026. Same protocol, same questions, same engines.
Measurement engine

What moved · August 9, 2026 → September 22, 2026

Eight indicators, measured twice the same way.

The assistants measured are ChatGPT, Claude, Gemini and Perplexity. Google AI Overviews is measured separately, on its own searches.

128 → 175

answers naming the brand

out of 1,000 answers per measurement

11 → 63

answers citing the brand on the 13 targeted questions

out of 260 answers · named or as a source

10th → 6th

rank among cited competitors

share of voice 3.29% → 4.42%

answers citing the assets produced

3 assets published · measured on September 22, 2026

outside top 10 → 6th

site's rank among Google AI Overviews sources

250 searches per measurement

26 → 35

Google AI Overviews answers citing the site

out of 250 searches

5th → 2nd

site's rank among all cited sources

4 assistants + Google AI Overviews

73 → 85

technical score out of 100

re-checked on September 6, 2026 · same pages

By assistant

Answers naming the brand, out of 250 per assistant.

August 9, 2026September 22, 2026

Perplexity 21 → 50

Claude 33 → 43

Gemini 57 → 65

ChatGPT 17 → 17

The published assets

54 answers cite at least one of the assets.

Pillar assetcited in 44 answers, 20 of which name the brand

Compliance assetcited in 11 answers, 4 of which name the brand

Videocited in no answer at this measurement

One answer cites the first two assets: it is counted only once in the total.

By question family

Where the brand gained ground, and where it lost ground.

Answers citing the brand (named or as a source), relative to the number of answers in each family. Families targeted by the action plan are flagged.

August 9, 2026September 22, 2026, upSeptember 22, 2026, down

Existing systemstargeted9 questions · 180 answers 4 → 47

Governance & trusttargeted4 questions · 80 answers 7 → 16

Use cases & automation3 questions · 60 answers 18 → 23

Category / discovery7 questions · 140 answers 15 → 24

AI-native & channels7 questions · 140 answers 35 → 41

Alternatives & comparisons8 questions · 160 answers 64 → 66

Identity & access (IAM/MDM)8 questions · 160 answers 66 → 59

What did not move

We publish that too.

ChatGPT: 17 answers out of 250 named the brand at both measurements.
Identity & access (IAM/MDM): 66 → 59 answers citing the brand, out of 160.

What these numbers say, and what they don't. Both measurements follow the same protocol; a gap below the noise floor (±15%) is not treated as a signal. We document what was published and what was observed afterward. We do not claim that one explains the other: other factors, such as model updates, can play a role.

Building blocks used

What was used in this case.

And in your market, what do AI engines answer?

Check my AI visibility → Contact us