A Digital Reality Check, end to end
A complete run of the framework on a fictional company, from the free Check through the Share of Voice tally to the monitoring plan. Follow it alongside your own Canvas.
Ridgeline is invented. The company, its competitors, and every number on this page are illustrative. Any resemblance to a real company is coincidental. The pattern is the point: this is what a zero-share-of-voice result looks like, and how to read it.
- The company: Ridgeline, a B2B platform selling grid-resilience planning software to mid-market utilities.
- The Check (free): the eleven queries in a clean session. 0 of 8 unbranded queries mentioned Ridgeline.
- The tally: 0 AI citations across all 7 platforms; the category leader had 228.
- The cause: not weak marketing. The internet cannot classify, resolve, or recommend the company. All three are fixable.
The three layers below are the Check, the Measure, and the Monitor, run in order. If you have already filled your Canvas, you can follow along and do the same for yourself.
The company
Ridgeline sells grid-resilience planning software. Its product joins two things most vendors keep apart: physical asset planning for substations, poles, and lines, and outage continuity modelling. Its core claim is that it is the only platform doing both in one place. It sells to mid-market utilities on six-figure annual contracts.
That makes AI representation a revenue question rather than a vanity one. Utility procurement teams now build vendor shortlists inside ChatGPT, Perplexity, and Copilot before they contact sales. A company with zero AI citations does not make the shortlist.
Layer 1: the Check (free, one afternoon)
We ran the eleven queries in a clean, logged-out session: eight unbranded and three branded, in two separate chats. Three of the unbranded ones:
Then the branded chat produced the tell that explains much of the rest. Because "ridgeline" is an ordinary geographic noun, query 9 returned a dictionary-style definition of the word rather than the company. The model cannot separate the brand from the generic term.
Result: 0 of 8 unbranded queries mentioned the company.
The three gaps behind it
These map onto three of the eight gaps, and each needs a different fix:
- Category & Positioning, plus Machine Understanding. AI reads "grid resilience" as cybersecurity and regulatory compliance. Ridgeline's real category, physical asset planning, is invisible, so it is never a candidate.
- Machine Understanding (entity resolution). The brand name collides with a common noun, so AI cannot resolve the company as a distinct entity.
- Value & Message, plus AI Representation. The "only platform doing both" claim has no software associations anywhere in AI. Nobody owns it, which makes it a wide-open category to claim.
Layer 2: the Measure, as AI Share of Voice
The Check shows what one platform does with specific queries. The tally shows the pattern across all of them. Counting citations for Ridgeline and four competitors is what dimensions 7 and 8 of the Digital Reality Score measure:
| Platform | Ridgeline | Comp. A | Comp. B | Comp. C | Comp. D |
|---|---|---|---|---|---|
| ChatGPT | 0 | 12 | 4 | 5 | 2 |
| Bing Copilot | 0 | 7 | 0 | 2 | 1 |
| Google Gemini | 0 | 13 | 2 | 3 | 0 |
| Google AI Mode | 0 | 41 | 8 | 9 | 5 |
| Google AI Overviews | 0 | 49 | 14 | 7 | 4 |
| Grok | 0 | 88 | 37 | 13 | 6 |
| Perplexity | 0 | 18 | 5 | 4 | 1 |
| Total | 0 | 228 | 70 | 43 | 19 |
Zero citations across all seven platforms. Not low, zero. The leader has 228, and even the fourth-placed competitor has 19. That single figure, AI Share of Voice: 0 across 7 platforms, benchmarked against named competitors, is the baseline the Score carries forward.
Layer 3: Monitor
A one-time reading is a snapshot. So the plan records ten monitoring prompts, each with the exact query, the platform to run it on, the specific competitors named in the baseline, a target, and how to verify. Quarterly, because AI answers move faster than search results. The 90-day target is deliberately modest and checkable: at least one citation on at least two platforms.
Closing the gap
The Check found the gap. Closing it means fixing the public evidence the machines actually read:
- Positioning: put the real category and the "only platform" claim where a first-time reader, and a model, meet them immediately.
- Structure: schema on the regulation and methodology posts, so answers can be extracted and cited.
- Entity: a definitions hub and an
llms.txtwith entity declarations, so the brand resolves as a company rather than a common noun. - Comparison and vertical pages, written to be citable rather than persuasive.
None of these are prompts or tricks. They make the real company provable, which is the only durable way to change what AI says about you.