When and How to Run an Assistant Round
“How do we know whether we appear in the assistant — there’s no panel. When and how do you run an assistant round?” No panel, correct — which is why measurement is manual and regular. Once a month, ten minutes, five questions, three notes. A one-off look misleads; a monthly record shows the trend. 📊
This article is the measurement method: the round’s rules, the record table and when to run it.
The whole line: the AI SEO guide; the local anchor: the anchor article.
The Round’s Rules
BU BÖLÜMÜN ÖZETİ
- Rule 1 — Five fixed questions
- Rule 2 — Same assistant, same form
- Rule 3 — Three notes
- Rule 4 — A clean session
The structure of the ten minutes: 🔍
Rule 1 — Five fixed questions
The same five questions every month: “psychologist in [district]”, “[main area of work] in [district]”, “should I see a psychologist or a psychiatrist”, “what happens in the first session” and “[clinic name]”. If the questions change, the comparison breaks: the questions article.
Rule 2 — Same assistant, same form
One assistant is chosen and used every month; if there’s a second assistant, a separate column. The question is typed in exactly the same words.
Rule 3 — Three notes
For each question: which sources were mentioned (platform, clinic, article), did our name appear (yes/no/wrong), what was said (one sentence). The third note is the most valuable — wrong information gets caught here.
Rule 4 — A clean session
Where possible, not from the clinic’s device but from a session with no history. If the assistant remembers previous chats the result gets personalised and the table breaks.
The Record Table
BU BÖLÜMÜN ÖZETİ
- The sixth row: mention count
- Side note: wrong information
- The second measure: “how did you find us”
One row is added to digital management’s five-row table: 📋
The sixth row: mention count
In how many of the five questions your name appeared: a single figure from 0 to 5. Recorded monthly, read quarterly — the same rhythm as the other five rows: the measurement article.
Side note: wrong information
If your name appeared but wrongly (old practitioner, old address, wrong title): a separate mark. This shows the source to fix that month: the wrong information article.
The second measure: “how did you find us”
The share of the answer “I asked the AI” in the front-desk record. The assistant round measures visibility; this answer measures conversion to bookings. The two are read together.
👉 The mention count tells visibility; “how did you find us” tells conversion.
When and How to Read It
BU BÖLÜMÜN ÖZETİ
- The baseline: before setup
- Monthly: the same day, the same time slot
- Quarterly: the trend reading
- And interpretive honesty
Timing and interpretation: 🗓️
The baseline: before setup
The first round is run before the four rungs are built. That’s the comparison point — the answer to “what changed” comes from here: the starting-point article.
Monthly: the same day, the same time slot
In the same week of the month, in daytime hours. Added to the routine’s monthly session as ten minutes; not a separate job.
Quarterly: the trend reading
Three months’ mention counts side by side: rising or flat. A single month’s result isn’t interpreted — assistant answers shift with time and wording; the trend emerges over three months.
And interpretive honesty
Five out of five isn’t the goal and isn’t realistic: intermediary platforms and national content are always ahead in some questions. A good picture for a local clinic: regular mentions in district questions and correct identification in the clinic-name question. If those two hold, the line is working. To inherit it built and district-locked: the parcel model.
📌 Field Notes
- When question wording is changed month to month the comparison breaks; five fixed questions are essential.
- Rounds run from the clinic’s own device give personalised results.
- Most wrong information is caught in the “clinic name” question: old practitioner, old address.
📖 Quick Glossary
- Assistant round: Monthly manual measurement with five fixed questions.
- Mention count: In how many of the five questions the clinic’s name appeared.
- Baseline: The comparison record taken before setup.
Frequently Asked Questions
➡️ Next Step
Write your five fixed questions this week and take the baseline. To inherit your district’s psychologist keywords built and locked, check your parcel; for timing, move to the timeline article.
Sık Sorulan Sorular
Interim action in two cases: when wrong information is seen (fixed that month) and when the mention count drops two months running (a contradiction has formed on some surface — the six-surface audit is run).
With four rules: five fixed questions (the same every month), the same assistant, the same words, three notes (source, did our name appear, what was said) and a clean session (no personalisation). Once a month, ten minutes.
A sixth row in the measurement table: mention count 0-5; as a side note, a wrong-information mark. The second measure is the share of “I asked the AI” in the front-desk record — mention count tells visibility, that answer tells conversion.
The first round before setup (the baseline), then monthly recording and quarterly reading. Alarm in two cases: wrong information and a two-month drop. The goal isn’t five out of five; it’s regular mentions in the district and correct identification on the clinic name.