A hyperrealistic illustration of what your report looks like — 12 prompts, 7 AI engines, 336 answers analyzed. Nothing here belongs to a real customer.
Five measured inputs, one weighted composite. The inputs are shown above; the exact weighting is GeoSignal's proprietary method — it's what makes the score hard to game and hard to copy.
Across 12 diner prompts and 7 AI engines (336 answers analyzed Sep 8 – Oct 6, 2026), Ember & Oak appears in 51% of AI answers — the strongest presence in your competitive set, and AI recommends you outright in 47% of those mentions. When Portland diners ask AI where to eat, your name comes up first and with the warmest language.
Your moat is specificity. “Best farm to table restaurant Portland,” “anniversary dinner Portland,” “wood fired pizza” — the prompts with a clear occasion or cuisine go to whoever has the clearest story, and your menu pages, chef interviews, and 2,100 Yelp reviews give AI quotable specifics. Generic prompts — “where to eat downtown,” “restaurants open late” — go to whoever has the most directory coverage, which today is Juniper Table.
The two prompts you've ceded are both logistics questions, not reputation questions. Nobody doubts the food; AI just can't find your hours, private-dining capacity, or late-night menu in a citable form. That's a fixable data problem, not a brand problem — and it's worth roughly 6,700 monthly searches.
Full sequencing with effort estimates lives under .
Same 7 engines, same prompt-level evidence, same 90-day plan — pointed at the brand you actually care about.