25 Questions to Ask Your CRM AI Assistant

Traditional CRMs organize data behind layers of navigation: Contacts tab, filter sidebar, custom report builder, export to CSV, pivot in Excel, present in meeting. Each question about your business requires learning where that answer lives in the interface. AI CRM assistants invert the model) you ask the question, and specialized agents find the answer by calling the right tools against live data.

Booked55's AI assistant routes requests to domain-specific sub-agents (contacts, companies, pipelines, tasks, events, campaigns, analytics, email, settings, and import) each with dozens of underlying API tools. You see one conversation; behind it, the system may chain multiple agent calls to answer a single question. This cookbook gives you 25 starting prompts across five categories. Adapt names, date ranges, and stage names to match your workspace.

Why Natural Language Beats Dashboard Clicking

Dashboards answer questions someone anticipated when building the report. Conversations answer the question you actually have right now (often one nobody pre-built a view for. “Which deals are stuck in Proposal with no email activity this month?” is a natural question. Building that report in a traditional CRM is a project.

Natural language also lowers the skill floor. New reps, solo founders, and part-time CRM users get the same analytical power as power users without learning filter syntax or report builders. The prompts below work in Booked55's built-in assistant and through MCP-connected clients like ChatGPT and Claude. For deeper context on the architecture, read what MCP means for CRM AI agents.

Contacts & Relationships (5 Questions)

1. Event-sourced contacts

“Find contacts from the Chicago Staffing Summit event.”

The events agent retrieves attendees linked to a specific event. Use this after conferences for segmented follow-up instead of manually tagging badge scans.

2. Top referrers

“Who referred the most clients? Rank by closed-won count.”

Combines contacts agent referral links with pipeline closed-won data. Identifies which relationships produce revenue) critical for referral pipeline building.

3. Duplicate detection

“Show duplicate contacts, same email or very similar names at the same company.”

Data hygiene without manual scanning. Follow up with merge or tag instructions once duplicates are identified.

4. Email history lookup

“What emails did we exchange with Sarah Chen? Summarize the last three threads.”

The email agent reads synced Gmail/Outlook threads. Essential pre-call prep and context before drafting follow-ups, see AI email generation with CRM context.

5. New contacts this month

“Show my new contacts this month, grouped by source.”

Quick pulse on prospecting output. Pairs with scoreboard data to verify activity translates to new relationships, not just busy work.

Pipeline & Deals (5 Questions)

6. Current pipeline value

“What's my pipeline value right now? Break down by stage.”

The pipelines agent returns live stage totals. Run this every morning as a 10-second forecast pulse (core workflow in our AI pipeline analysis guide.

7. Stuck deals by stage

“Which deals are stuck in Qualification for more than 14 days?”

Stage duration filtering without custom reports. Replace [Qualification] with any stage name in your pipeline configuration.

8. Move a deal

“Move the Acme Corp deal to Proposal stage.”

Write operations through conversation. Update stages, values, and deal details without opening the board, useful for mobile updates via MCP.

9. Funnel conversion

“Show funnel conversion for Q1, how many deals entered each stage and how many closed won?”

Analytics and pipelines agents combine for pass-through analysis. Reveals where deals drop off in your sales process.

10. Stage duration benchmarks

“How long do deals sit in Qualification on average? Compare to last quarter.”

Historical stage duration trends inform coaching and forecast realism. Longer averages in early stages may signal qualification criteria issues.

Analytics & Performance (5 Questions)

11. Scoreboard comparison

“Compare my scoreboard to the team average this month.”

The analytics agent pulls scoreboard metrics, contacts, tasks, events, activities, deals, for individual and team views. Objective input for coaching conversations.

12. Activity breakdown

“Show my activity breakdown this month, contacts created, tasks completed, deals moved, events attended.”

Understand where effort goes before asking why results lag. Connects to measuring sales activity to grow revenue.

13. Close rate by rep

“Which rep has the highest close rate this quarter? Show deals closed vs. deals created.”

Team performance visibility without building custom dashboards. Useful for managers running small teams without dedicated ops support.

14. Contacts created trend

“Chart my contacts created by month for the last six months.”

Prospecting consistency over time. Declining new contact creation often precedes pipeline gaps two to three months later.

15. Referral conversion

“What's our referral conversion rate, referred contacts that became clients vs. total referred?”

Quantifies referral program effectiveness. Low conversion may indicate poor referral qualification or weak follow-up on introductions.

Tasks & Follow-ups (5 Questions)

16. Overdue tasks

“What tasks are overdue? Sort by linked deal value.”

Daily essential. Overdue tasks on high-value deals are the highest-priority recovery actions, see finding clients who need follow-up.

17. Create a follow-up task

“Create a follow-up task for John Martinez on Friday (call to discuss proposal timeline.”

Task creation through conversation. Links to contacts and deals automatically when names resolve in CRM context.

18. This week's tasks

“Show tasks due this week, grouped by day.”

Weekly planning in one query. Replaces opening the task board and manually scanning due dates.

19. Task completion rate

“What's my task completion rate this month) completed vs. created?”

Discipline metric. High task creation with low completion suggests over-planning; low creation suggests reactive rather than proactive selling.

20. Bulk follow-ups by stage

“Bulk assign follow-up tasks for all contacts in Qualified stage (due next Wednesday, task type: call.”

Pipeline hygiene at scale. One request creates tasks for an entire stage segment instead of clicking each contact individually.

Events & Campaigns (5 Questions)

21. Event attendance

“Who attended the Annual Advisor Conference? Show contacts not yet followed up.”

Post-event action lists. Cross-reference attendance with task and email activity to find gaps in event ROI capture.

22. Nurture campaign creation

“Create a nurture campaign for contacts tagged Healthcare who have not been emailed in 90 days.”

The campaigns agent builds audience segments from contact criteria. Refine copy in Unlayer before launch.

23. Launch a campaign

“Launch the Q1 re-engagement campaign.”

Operational campaign management through chat. Confirm audience size and schedule before executing send commands.

24. Mark attendance

“Mark Lisa Park as attended for yesterday's webinar.”

Quick event logging during or after events. Keeps attendance data current for post-event segmentation queries.

25. Event invitation email

“Draft an event invitation email for contacts in the Boston metro area who attended last year's event.”

Combines contacts segmentation, event history, and email agent drafting. Review and send through connected Gmail or Outlook.

Tips for Effective AI Queries

These prompts work better with a few habits:

• Be specific with names and stages) “John Martinez” resolves better than “the Acme contact.” Use exact stage names from your pipeline.
• Include date ranges(“this month,” “last 30 days,” “Q1 2026” anchor time-bound questions.
• Chain analysis and action) Follow “show stale deals” with “create follow-up tasks for each” in the same conversation.
• Challenge unexpected results, Ask “why is this contact included?” to understand AI reasoning and catch data hygiene issues.
• Save recurring queries as habits, Same three morning questions beat inventing new ones daily.

For solo reps building a daily routine, see AI CRM for solo sales reps. For MCP setup outside Booked55, see using Claude with CRM via MCP.

Booked55: 10 Specialized Sub-Agents, 90+ Tools

Every prompt above routes through Booked55's sub-agent architecture. The parent AI assistant sees high-level domain tools; each sub-agent executes with specialized API access:

• contacts_agent, companies_agent, pipelines_agent
• tasks_agent, events_agent, campaigns_agent
• analytics_agent, email_agent, settings_agent, import_agent

Combined, these agents expose 90+ CRM operations (search, create, update, analyze, send) without you learning API syntax. Natural language is the interface; agents and tools are the engine. This is what AI-native CRM architecture looks like in daily use: not a chatbot bolted onto forms, but intelligence woven through every data domain.

Building Your Personal Query Playbook

The 25 prompts above are starting points. Mature CRM AI users develop a personal playbook, five to ten queries they run on fixed cadences:

• Daily (90 seconds), Pipeline value by stage, overdue tasks, contacts with no activity in 14 days
• Weekly (5 minutes) (Stale deals by stage, new contacts by source, task completion rate
• Monthly (10 minutes)) Funnel conversion, referral metrics, ICP pattern refresh on closed-won deals

Document your playbook in a note doc or pin queries in your MCP client. Consistency matters more than novelty, the rep who runs the same three morning questions daily outperforms the one who invents new prompts weekly but skips half the sessions. Adapt stage names and tag conventions to your workspace once; reuse forever.

When the AI Gets It Wrong

AI queries occasionally return unexpected results, ambiguous contact names, outdated tags, or misinterpreted date ranges. Recovery is conversational:

• “I meant John at Acme Corp, not John at TechStart (redo the query.”
• “Exclude closed-lost deals from that pipeline summary.”
• “Why did you include this contact? Show me their tags and last activity.”

Treat unexpected results as data hygiene signals. If the AI includes contacts you consider inactive, your tags or stage definitions may need cleanup, a worthwhile side benefit of regular querying.

Pricing and Getting Started

All 25 query types work on every Booked55 plan (AI assistant, MCP server, enrichment, campaigns, and analytics included. $129/month for the first seat plus $59/month per additional user. Pick three prompts from this list, run them tomorrow, and expand from there. Try the AI assistant at booked55.com.

The Bottom Line

• Start with five prompts in one category) Master contacts or pipeline queries before expanding to campaigns and events.
• Be specific with names, stages, and dates, Precision in the question produces precision in the answer.
• Chain read queries with write actions, Analysis without follow-up task creation leaves value on the table.
• Run the same morning trio daily (Pipeline value, stale deals, overdue tasks) 90 seconds that replace a 30-minute board review.