Visitor Screening
Teach your Aijent who is not a fit, so it stops trying to sell to people you cannot help.
Visitor Screening is where you describe who is not a fit, so your Aijent stops trying to convert them. It is on every plan and off until you turn it on. You write two plain-English answers, your Aijent turns them into the signals it watches for, and every conversation then gets a fit rating you can see in your analytics.
Not everybody who lands on your website is somebody you want to meet. Visitor Screening is where you describe who is not a fit, so your Aijent stops working to convert them.
It is on every plan, and it is off until you turn it on.
You may have already written it
The setup wizard asked who your ideal customer is, and who your Aijent should politely screen out. If you answered the second question, most of this is already done.
If you skipped it, your Aijent currently treats every visitor as a potential customer.
Setting it up
On the Qualification tab, you write two plain-English answers:
Who is your ideal customer?
Who should your Aijent politely screen out?
Then select Generate criteria. Your Aijent turns those answers into the signals it will actually watch for, split into good fit and screen out. Review them, then save.
Change your answers and regenerate whenever your business changes.
Writing answers that work
Be specific and be concrete. "Small businesses" is not something anybody can judge. "Businesses under ten staff, or anyone outside Australia and New Zealand" is.
For who to screen out, think about the enquiries that waste your week: the wrong size, the wrong place, wanting something you do not offer, and the people who were never buyers at all, such as students, job seekers and competitors.
Screen on who someone IS, not how they are behaving
Screening works on lasting traits, not on how a conversation happens to be going. "Just browsing" or "has not mentioned budget yet" are not reasons to screen somebody out: plenty of serious buyers start exactly like that. Your Aijent deliberately ignores that sort of instruction.
When it is unsure
There is a setting for conversations your Aijent cannot call either way.
Carry on and engage anyway is the recommendation. If in doubt, talk to the person. It protects you from turning away business over a bad guess.
Route to a human hands it to your team instead of pushing on.
This setting also governs what a SureGate Review does, so changing it changes both.
What you see afterwards
Every conversation gets a rating, shown as Fit in your analytics:
- Good fit
- Below fit
- Not a fit
You can filter on it, and each conversation shows the reason. That is worth reading for the first week: it tells you whether your screening is catching what you meant, or turning away people you actually wanted.
A very short conversation, where the visitor barely engaged, is marked as not a fit without any real judgement being made. Do not read too much into those.
How this differs from SureGate
Visitor Screening is a judgement. It weighs everything it knows and forms a view, the way a person would.
SureGate is a rule. Exact numbers, enforced by the server, with no judgement involved.
Most businesses want both: screening for the fuzzy sense of a poor fit, SureGate for the one or two lines that are genuinely absolute. SureGate needs screening switched on first, so this is where you start.
Support Aijents never screen anybody, since their job is to help rather than qualify.