About Emphanic
About Emphanic
Emphanic is built for one thing: measurable visibility in AI answers for small and medium businesses.
Behind the product sits experience from web, product, SaaS, startups and AI. That combination is the point. AI visibility is not a marketing trick and not a purely technical problem, but both at once: measuring takes engineering, interpreting takes product thinking, and the fixes have to be workable for an owner without a technical team.
We keep the team small and the automation large. The measuring runs as a system, every week, in the same order. People do what people are better at: designing the measurement, weighing the outcome and answering your question. No account managers, no layer in between.
Every measurement
- 4
- AI assistants
- 12
- questions per assistant
- 48
- answers per measurement
- 7
- days to the next one
The full measurement design is in the methodology. Read the methodology
01 / THE WEEKLY MEASUREMENT
What runs every week
Six steps, every week, in the same order. The system does the work; the design and the judgement around it are human.
Check the access
What the system does
Reads your robots.txt, your llms.txt and your sitemap, and checks whether the crawlers from OpenAI, Anthropic, Perplexity and Google are allowed through.
What you see from it
Which door is shut, and the line that opens it.
Ask the questions
What the system does
Puts twelve real customer questions per assistant to ChatGPT, Claude, Perplexity and Gemini, with web access on and in your region.
What you see from it
Forty-eight answers, with the question and the date attached.
Take the answers apart
What the system does
Looks for your business name and its spelling variants in every answer, and extracts the sources the assistant cites.
What you see from it
Your score per assistant, with the denominator next to it.
Add the competition
What the system does
Collects which other businesses get recommended in the same answers, and how often.
What you see from it
Who is standing in the spot where you are not.
Write the fixes
What the system does
Writes the structured data that is missing: schema.org JSON-LD, question and answer blocks, and rules for your robots.txt.
What you see from it
Files you copy, or one line of code that puts them live.
The judgement call
What the system does
A person looks at the outliers, calibrates the questions and answers what you ask.
What you see from it
A report that holds up, and an answer to what you asked.
Then it starts again. Whatever you changed goes into the next measurement, so you see for yourself whether it worked.
02 / CHOICES
Why it works this way and not another
Five choices that separate a measurement from a sales pitch.
01
Weekly, not daily
AI answers shift over weeks, not days. Measuring more often mostly produces noise that reads like a trend.
02
Twelve questions, and the denominator stays
A single question says nothing about a system that does not answer the same way twice. Twelve per assistant gives you a denominator, and that denominator stays wherever the number goes.
03
With web access on
We query the assistants the way your customer uses them. A model without web access answers from memory, and that is not what appears on your customer's screen.
04
A flat week is a result too
We report what we measure, including the weeks where nothing improved, and every source stays visible. A service that only sends good news is not measuring anything.
05
No promise about your place in ChatGPT
Nobody can guarantee that, us included. Anyone who promises it is selling something they cannot deliver.
What you are risking
- Seven days free, and only then the first charge
- Cancel monthly, with one click through Stripe
- Every schema and every file we deliver stays yours
- No VAT added on top of the amount on the pricing page
Mission
Every business that does good work deserves to be found, now that the search bar has become a conversation. We make AI visibility measurable, affordable and actionable for businesses without a marketing department.
Vision
Within a few years, the search for a local business will start at an AI assistant. A business that is not named there stops existing commercially. We want an owner to be able to follow and adjust their position in that answer as routinely as their revenue.
03 / THE WORK
Kept small, on purpose
The measuring is repetitive and has to happen exactly the same way every week, so a machine does it. Whatever takes judgement is done by a person. That is what keeps a weekly measurement sustainable without losing anything in quality.
People
- 01Designing the measurement: which questions, which region, which assistants
- 02Adjusting the questions when an assistant changes its behaviour
- 03Weighing the outliers before a report goes out
- 04Answering your mail
The system
- 01Putting the same questions to four assistants every week
- 02Checking whether the crawlers can reach your site
- 03Taking every answer apart for your name, its variants and its sources
- 04Writing the structured data that is missing
Emphanic is built and run by Tom Suter, from Appenzell. No ticket system and no account manager: your question lands with the person who built the system, and that is where the answer comes from too.
- Address
- Haslenstrasse 9, 9050 Appenzell, Switzerland
- Phone
- +41 76 4 123 242
- info@emphanic.com
The name comes from the Greek 'emphanes': visible, clearly present.
Measure your AI visibility
The check is free and returns results within 90 seconds.