How it works
From measurement to report, every step is recorded.
- Setup ≈ 2 min
- First scan within 24h
- 6 components · 1 score
- Unmeasured ≠ zero
Four steps from scan to report
For every step: what happens, what you see and how long it takes.
- 1
Setup and business card
- What happens
- Add name, address, category and website. Brand spelling variants and competitors are added to the list; Ajans Pro manages up to 10 brands in one panel.
- What you see
- One card with the business profile, service area and competitor list — plus a warning for any missing field.
- How long
- ≈ 2 minutes. Setup fee applies only to Yönetim: ₺1,500 in 3 installments.
- 2
Query set and lenses
- What happens
- 30-40 queries are built from region, category and intent variants, each tagged with a lens: geography, price, persona, time, comparison, trust.
- What you see
- The query list in the panel, ready to approve; the set is versioned so next month compares like with like.
- How long
- Built automatically; first approval ≈ 5 minutes.
- 3
Measurement: engine calls
- What happens
- The same set is sent to real AI engines. Every answer stores mention, rank, cited domains, model + version, date, location and cost. Free-user and paid-user profiles are twin-scanned.
- What you see
- A live view: queries run, mentions found, rank per engine — with the raw answer text underneath.
- How long
- First pass within 24 hours. Daily light scan every day; weekly grounded pass on Üst and Ajans Pro.
- 4
Report and action plan
- What happens
- The score is split into components; competitors and citations are tabulated. Trigger rules produce tasks and rank them by impact × effort.
- What you see
- PDF, mobile link and WhatsApp summary, with 'done this month' kept apart from 'recommended'.
- How long
- The month-end report is drafted automatically; the score impact is measured 4 weeks later.
Six components, one score
Nothing hidden: each component's weight, definition and refresh cadence is below.
| Component | What we measure | Weight | Refresh |
|---|---|---|---|
| AI mentions | In how many of 30-40 queries your brand appears, including variants and citations. | 40% | Every scan |
| Rank quality | Your position in answers and how often you land top three. | 20% | Every scan |
| Source power | Cited domains and your share of those sources. | 15% | Weekly |
| Google Maps | Local pack position, rating and review count. | 15% | Weekly (with SERP key) |
| Social trace | Whether your social profiles appear in AI answers. | 10% | Weekly (semi-manual) |
| Verification | Twin-scan agreement and manual calibration result. | Confidence | Monthly |
Scroll the table sideways.
Weights are fixed and printed in the report version. Unmeasured components never enter the score — coverage drops instead. The five weighted components total 100%.
Which engines?
Measurement runs on the same model APIs consumers use. No scraping; model and version are logged for every answer.
- OpenAI / ChatGPT
- Grounded web search and model memory are measured separately, with user_location set to your area. No VPN or browser farms.
- Gemini
- Grounded layer opens on paid tiers; without web search on the free tier that measurement is labelled 'memory' in the report.
- Compatible providers
- Perplexity and Claude-class APIs. The report names the engine on every query row.
Weekly and monthly rhythm
Cadence follows the plan; the methodology does not. The query set is re-runnable, so month-to-month comparison stays meaningful.
- DailyLight scan: memory and trend. Included in İzleme.
- WeeklyGrounded pass: OpenAI + Gemini + Perplexity, on Üst and Ajans Pro.
- MonthlyFull report plus manual calibration: 3-5 queries checked in the consumer app.
- MonthlyAgreement rate: how well the API measurement represents the real experience.
What is in the report?
A decision tool, not decoration. Every report follows the same skeleton, so months compare cleanly.
Executive summary
Three sentences: score, movement since last month, the single most important job.
Query table
Engine, mention, rank and a verbatim quote for every query.
Competitor benchmark
Who appears in the same answers, how often, and your share.
Citations and sources
The domains AI cites and your site's share of them.
Action plan
3-7 tasks scored by impact/effort, each with steps, owner and time estimate.
PDF, mobile and public link
Downloadable PDF, mobile reading link, WhatsApp summary and an optional public report link.
Every report opens with the coverage percentage and a verification box: scans run, components measured, agreement rate.
Honest limits
A measurement that hides its limits is not a measurement. These are printed in the report on purpose:
- AI answers change within a day. A single scan is not absolute truth, so trend, repeatability and coverage are shown.
- Unmeasured components are not counted as zero; the report says 'not measured'.
- Google Maps needs a SERP provider key. Without it the component is off, and the report says so.
- Grounding differs per engine: Gemini's free tier has no web search, so that figure is labelled memory.
- Below 20 samples we publish a ratio and an 'early signal' tag, not a raw count.
- The score is not a ranking guarantee. Our job is to measure, prove and point the next step.
Frequently asked
Is this an SEO tool?⌄
No. It measures presence in AI assistant answers, not classic search rankings. Google Maps is a separate component.
What if the score is 0?⌄
That is a measurement: you are not mentioned yet. The report explains why and gives the first three steps — it never invents a score.
How often does a report arrive?⌄
1, 2 or 4 per plan: month-end full report on İzleme; mid-month glance plus month-end on Yönetim; weekly pulse plus month-end on Üst.
Can I verify the measurement myself?⌄
Yes. Every row carries the raw answer, model, date and sources. Run the same query in your own ChatGPT or Gemini account and compare; differences are already reported as 'tier difference'.
Where is my data kept, and for how long?⌄
In the EU (Frankfurt) with tenant-level row security. Raw data stays active for 24 months, then archives; no personal data is processed and everything exports to JSON/CSV.
See the free preview first
Business name, location and your name are enough. We measure mentions and maps position with real AI queries and show it immediately.
No credit card · 15-second preview · The result is a preview; the full score lives in the panel.