No black box - a measured, five-step process with a baseline captured before we touch anything.
What you get
Entity coverage
Percentage of your business's canonical surface (name, locations, services, providers) that is unambiguously modeled in every AI-facing surface.
Citation quality
Number and authority-weight of sources citing your business, tuned to each target model's preferred source stack.
Share of voice (SOV)
Of a curated set of 30–100 buyer prompts, the percentage where your business is named in the AI answer.
Durability score
Internal grade for how likely each signal is to survive the next model update. We only ship high-durability work.
How it works step by step.
Prompt inventory
Curate 30–100 real buyer prompts across category, comparison, geography, and use-case dimensions.
Baseline scoring
Run every prompt through 6+ major AI models. Score SOV, citation composition, and entity accuracy.
Monthly re-scoring
Same prompts, same models, every month. Track deltas and publish side-by-side screenshots.
Quarterly recalibration
Refresh the prompt inventory as buyer language evolves and models ship new capabilities.
Frequently asked questions
Why not just track rankings like SEO?+
AI answers aren't ranked - they're synthesized. A prompt either names your business or doesn't. Share of voice is the honest analog.
How many prompts do you track?+
Typically 30–100 per engagement. Enough to be statistically stable, small enough that we manually review every screenshot each month.
Do you use APIs or scrape?+
We use public interfaces where models offer them, and official APIs where they don't restrict measurement use. No ToS-breaking scraping.
Can we see the raw data?+
Yes. Every screenshot, every prompt, every citation source lives in your shared workspace.
Ready to get recommended by AI?
Allocation is capped. Apply now to see if your business qualifies.