Sales Assistant Playbook

Prompts and patterns that qualify leads, book calls, and fill your pipeline

The most common reason a sales assistant underperforms isn’t the model — it’s the prompt. This playbook covers the four moves that turn a chatbot into a sales rep: qualify, hand off, track, and follow up.

1. Qualify every visitor

A good sales assistant asks four questions on the first conversation:

1. Project — what they need help with
2. Budget — rough price range they’re working with
3. Timeline — when they want to start
4. Decision-maker — who else is involved

Sample agent instruction

You are a sales assistant for {company}. When a prospect engages, ask:

1. What's the project they need help with?
2. What's their budget range? (under $5k / $5–25k / $25k+)
3. What's their timeline? (Now / This month / Later)
4. Who else is involved in the decision?

Save their answers to the CRM. If they're a fit (budget $5k+, timeline
Now or This month), offer to book a discovery call via your calendar.
If not a fit, politely thank them and suggest reaching out later.

Use the Readiness custom field

The coaching_consulting and similar CRM templates ship a Readiness field with three options: Now, This month, Later. Have the agent ask the timeline question and save the answer to this field — not free text.

This makes filtering trivial later: “show me every prospect with Readiness = Now and Budget > $5k.”

Every CRM template ships with custom fields tuned for that go-to-market shape. See the CRM templates reference for the full list of fields per template.

2. Hand off to a calendar

Once a prospect is qualified, don’t ask “what time works?” — let your calendar tool find slots. Connect Calendly or Cal.com so the agent can suggest slots and book directly from the chat.

Sample handoff prompt

When a prospect is qualified and asks to book a call, use the
list_event_types tool to find available event types, then use
get_availability to find slots in the next 7 days, and create_booking
to lock in their preferred time. Confirm in chat with the booking
link.

The agent does the back-and-forth (which slot? which timezone? which event type?) so the prospect doesn’t bounce to a separate scheduling page.

3. Set up the right CRM pipeline

BoostGPT ships 5 pre-built CRM templates — pick the one that matches your sales motion. The right template means you don’t start from an empty board.

Template	For
agency_client_intake	Agencies turning inquiries into discovery calls + signed contracts
coaching_consulting	Solo consultants/coaches moving applicants to enrollment
sales_pipeline	General B2B — qualify, proposal, close
support_requests	Service-request tracking (not for new lead gen)
simple_lead_tracker	Lean fallback if none of the above fit

Pre-built stages (agency_client_intake)

New Inquiry → Discovery Call → Strategy Sent → Contract Sent → Closed

Each template includes contact statuses, pipeline stages, and structured custom fields tuned for that motion. The agent saves qualified leads directly into this pipeline — no manual data entry.

4. Automate follow-ups

The biggest reason leads go cold isn’t the lead — it’s slow follow-up. Activate workflows so nothing slips:

Starter workflow (auto-created at signup)

When you finish onboarding, BoostGPT creates a draft “Notify me on new leads” workflow:

• Trigger: crm.contact.created — fires when a lead is captured
• Action: notify_owner — sends you an email or in-app notification

It ships in draft state. Activate it from the dashboard’s amber callout (one click) or from the agent’s Workflows tab.

Recommended additions

Workflow	Trigger	Action
Follow up if no reply	Deal moves to “Proposal Sent”	Wait 3 days → create CRM follow-up task
Hot lead alert	New lead with Readiness = “Now”	Notify owner immediately (urgent priority)
Won deal handoff	Deal stage changes to “Won”	Create CRM note + notify owner

See the Workflows tab on each agent for the full template library.

Best practices

Don’t push prospects who aren’t a fit. A polite “we don’t seem to be the right fit right now” beats an annoyed prospect telling friends not to use you.

Train on your services + pricing + FAQ pages. The agent answers like you do only if it’s read what you’ve written. Add your URL during onboarding or in the Training tab.

Use specific numbers. “$5k–$50k engagements” is sharper than “service businesses.” Concrete numbers in your training data help the agent qualify accurately.

Test qualification questions. The Insights tab shows which conversations led to bookings. Iterate the prompt based on real conversations, not guesses.

Next Steps

Calendar Integration

Connect Calendly, Cal.com, or Google Calendar

CRM

Custom fields, pipeline stages, deals, tasks