Quincer AI Docs Home Support Start free

Configure

Sales automation

Quincer AI doesn't just answer questions. It qualifies every visitor against your BANT profile, books meetings with the right rep while intent is hot, and keeps the conversation alive by email afterward. This page covers how each piece works and how to tune it.

BANT lead scoring

Every conversation starts at 0 and accumulates points as the visitor shares signals. The formula is additive, with four dimensions:

DimensionSignalPoints
BudgetVisitor mentions a budget range that meets or exceeds your floor+25
Visitor confirms budget is available (but no number)+10
Visitor has decision-making authority over budget+5
AuthorityJob title indicates decision-maker (VP, Director, Head, Founder, CXO)+20
Title indicates influencer (Manager, Lead, Senior IC)+10
NeedStated pain matches your ICP pain description+20
Evaluating solutions (compared to competitors by name)+10
TimelineBuying within 30 days+25
Buying within 90 days+10
No timeline mentioned0

Tiers

ScoreTierDefault action
70-100HotOffer meeting link + Slack alert to sales.
40-69WarmAsk one more qualifying question, then offer content + follow-up email.
0-39ColdCapture email for nurture; do not alert sales.

Adjust thresholds and per-signal weights per persona from Personas → [persona] → BANT profile.

Meeting routing

Once a lead crosses your Hot threshold (or whenever they explicitly ask for a demo), Quincer AI offers available times and books the meeting on the right rep's calendar. Routing follows this order:

  1. Territory rules first. Match on country, state/region, or company size.
  2. Round-robin within the matching rep pool, skipping reps who are OOO.
  3. Fallback to the team's default calendar if no territory matches.

Configure territories under Organization → People (the Territories controls). You can scope rules by any combination of: country, region, employee count, industry, visitor URL, or persona.

Calendar sources

Meetings are booked via whichever calendar integration is connected:

If more than one is connected, Quincer AI prefers the calendar explicitly assigned to the routed rep.

Email follow-up

When a conversation ends (visitor closes the widget or 30 minutes of inactivity), Quincer AI can send a follow-up email. You set this under Customize & Deploy → Email Configuration. The panel below shows that screen: a status banner at the top confirms which mailbox is the active sender (or warns that none is connected), and a single Email follow-up mode dropdown controls what the agent does after each chat.

Email Configuration
Sending mailbox

Emails are sent from sarah@company.com (Gmail). Visitor follow-up emails are sent from this address.

Email follow-up mode
Review — Agent drafts emails, you approve before sending

Other option: Auto — Agent sends emails immediately (Scale plan or the email follow-up add-on). With no mailbox connected, follow-ups won’t send regardless of this setting.

The Email follow-up mode dropdown — Review drafts for your approval, Auto sends immediately.

There are two modes in that dropdown. Review — the agent drafts the email and you approve before it sends — is the default. Auto — the agent sends immediately — requires the Scale plan or the email follow-up add-on, so it's disabled otherwise. Either way, the email is personalized with the visitor's name, the specific pain they described, any tools they mentioned, and a clear next step (meeting link, content, or reply ask).

Requires a connected mailbox. Follow-up emails go out only from your brand's connected Gmail or Microsoft 365 account — the sending mailbox shown in the banner above — so the visitor can reply to a real, monitored inbox. If no mailbox is connected, the banner turns amber and no follow-up is sent (Quincer never emails visitors from its own address).

Cadence

Multi-touch follow-up is available on Growth and Scale. You can configure up to 5 touches over 21 days, with branching based on whether the lead replied or booked.

Lead notifications

Get pinged the moment the assistant captures a lead — their name, email, and what they asked for — instead of checking the dashboard. Configure it under Leads → Lead notifications. In the card below, each recipient is a removable chip; type an address in the name@example.com box and click Add, then Save. The hint at the bottom tells you which mailbox the ping is sent from.

Lead notifications

Email these people the moment the assistant captures a lead (name, email, and what they asked for). They don’t need a Quincer login.

frontdesk@clinic.com × sarah@company.com ×
name@example.com
Add
Save Saved ✓
☐ Also attach each lead as ADF XML ⓘ

Sent from your connected mailbox if you’ve linked one (Integrations → Email) — otherwise from Quincer.

The Lead notifications card — add any email (no login needed), then Save. Fires once per lead.

ADF for automotive CRMs

If your CRM has asked for ADF — Auto-lead Data Format — tick Also attach each lead as ADF XML in the same card. Each notification then carries a small .xml file that dealer CRMs read automatically, so nobody retypes the lead by hand.

i

Leave this off unless someone asked for it. ADF is an automotive industry standard. Outside that world it is just an extra attachment nobody opens. The email itself is identical either way — ticking the box only adds a file.

What we put in it:

!

No vehicle block. ADF expects a year, make and model, and a Quincer lead does not capture those — so we leave the vehicle section out rather than invent one. Most CRMs accept the lead without it; if yours rejects it, tell us and we will look at capturing vehicle interest in the chat.

A lead with no name, or with neither an email nor a phone number, cannot form a valid ADF record. Those still arrive as a normal notification — just without the attachment.

Tip for appointment-style businesses (clinics, salons): the assistant can't book into a non-Google calendar, so the reliable flow is to capture the request and notify a team member here, then a human books it — or use a Calendly link for self-serve scheduling.

Escalation to humans

The agent hands off to a human when any of these fire:

Where that handoff lands is set per persona, under Personas → [persona] → Channels in the Where your team gets notified group — one row for each messaging platform your brand has already connected (Slack, Microsoft Teams, Google Chat, Telegram), each collapsed to a one-line summary of what it's currently set to. A platform you haven't connected gets no row here at all: it sits in the dashed Not set up yet group further down the tab as one line with a single Connect button, which sends you to Integrations — never a greyed-out form. So if you only run Slack, Slack is the only destination row you'll see. Open the Slack row and pick a channel under Escalations go to: the persona posts each escalation there and team members reply in the thread. Leave it on None — escalation off and escalation is disabled for that persona. Live activity goes to is a second, optional channel that keeps proactive pings out of the escalation channel — if you leave it unset, they fall back to the escalation channel.

Where your team gets notified
Outbound. Escalations, live pings and summaries — one row per destination.
Slack
Escalations → #sales-escalations · live → #live-chats
• Active
Escalations go to
# sales-escalations

Team members reply in the thread.

Live activity goes to
# live-chats

Keeps proactive pings out of the escalation channel. If unset, falls back to it.

Send these to your team

New conversations and qualified leads post to the live channel; the end-of-conversation summary posts to your conversation-history channel.

✓ New conversations ✓ Qualified leads Conversation summary
Escalate on negative sentiment
Off — negative conversations are not auto-escalated
Off

The expanded Slack row — Escalations go to picks the channel, or None — escalation off to turn escalation off for that persona.

Every destination asks the same questions with the same labels: Microsoft Teams picks a channel, Google Chat picks a space, and Telegram takes an Escalation chat ID you type in instead of a picker (it has no separate live target). Two fields are extras: [Persona] can DM — the teammates this persona may message directly, on Slack and Microsoft Teams — and Assistant channel, Slack only, where your team can chat with the persona itself; a plain post or an @mention is answered in-thread using its knowledge and tools, and counts as a conversation.

Send these to your team is the chip row underneath. New conversations posts the moment a visitor engages, with a Take over button so someone can jump in before the AI escalates; Qualified leads posts when a lead first provides an email or their score crosses the qualified threshold. Both go to the live channel. Conversation summary posts an end-of-conversation summary to the conversation-history channel or space you set under Integrations — not the live channel. Telegram has no chips at all: there is no per-persona Telegram ping or summary setting, so a Telegram summary only happens when the agent decides to call its own summary tool.

Rows open one at a time, and a row is collapsed until you click it — with one exception: when the tab loads, a row that needs fixing opens itself, because a problem nobody sees is a problem nobody fixes. Five rows can claim that slot — Slack, Microsoft Teams, Google Chat, Escalate on negative sentiment and Phone — and the first one with a problem, in that order, is the one that opens. From your first click onward you control which row is open.

An amber Needs attention pill is what marks those rows. It fires when a saved channel or space is no longer listed on the platform, when the agent was never invited to the Slack channel it's meant to post in, when the platform rejected our request for its channel list, when a Microsoft Teams or Google Chat link is still half-finished (the app hasn't been added in Teams, or the Google Chat space claim is still pending its connect code), when Escalate on negative sentiment is on with no target chosen — or with a target whose own Escalations go to is empty, so the escalation would fire into nothing — or when a phone number assigned to this persona is switched off, so calls to it go unanswered. One amber pill deliberately opens nothing: the Email inbox line in Not set up yet, shown when mailbox monitoring is on but no mailbox is connected. There is no row to expand in that state, and the warning is already on screen.

The last row in the group, Escalate on negative sentiment, escalates on its own when a visitor sends several negative messages in a row. Pick where it lands under Escalate to — a platform's Escalations go to target from the rows above, or Flag for a human, which marks the conversation on the Inbox → Live view — then set the negative-streak threshold and a cooldown. It needs the Scale plan, with sentiment scoring on under Analytics → Sentiment.

See Live chat for how the human takeover works.

Where to tune it