Use cases

Prerequisites

1. Connect your account: Connect your account if you use a managed AI client.
2. Build a custom client: Configure authentication if you are building a custom client.

Use these example prompts and workflows after connecting to Mindtickle MCP. For what each tool does and the scope it requires, see Supported tools.

The examples fall into three groups: single-tool examples, where one prompt maps to one tool call; multi-tool examples, where your AI client chains several tools together to complete a task; and patterns for developers building these workflows into their own client. You don’t need to name a tool in a prompt. Your AI client picks the tools based on what you ask.

Each example works only if your connection includes the scopes for the tools it uses. If your AI client can’t find a tool, check which scopes were granted. The Manager Insights prompts need the read:manager scope, and they work only for a user who has the manager role in Mindtickle. The prompts about your own training progress need no scope.

How the tools connect

Most multi-tool workflows follow one rule: one tool returns an ID, and the next tool takes it. Knowing which ID goes where explains why your AI client calls tools in a particular order.

ID	Returned by	Used by
company_id	find_companies, find_deals, find_rooms, find_contacts	find_contacts, find_deals (as company_ids), create_room, edit_room
deal_id	find_deals, find_rooms	create_room (required for deal rooms), edit_room
deal_guide_id	find_deals	fetch_deal_guide
contact_id	find_contacts, create_contacts	create_room and edit_room (as contact_ids)
Contact email address	find_contacts, create_contacts	share_room (as invitees)
room_id (UUID)	find_rooms, find_deals, create_room, edit_room, fetch_deal_guide	fetch_room
room_id_int (numeric)	find_rooms, find_deals, create_room, fetch_room, edit_room	edit_room, share_room
Asset id	mindtickle_search	create_room (as asset_ids), edit_room (in assets_to_add)
Call id	find_call (as id), fetch_deal_guide activity timeline (as call_id)	fetch_call (as callIds)
Tutorial tutorial_id	find_tutorial (as tutorial_id)	fetch_tutorial (as tutorialId)
Series series_id	fetch_manager_team_insights (as series_id in the series list)	fetch_manager_team_insights (as object_id, with object_type set to series)
Module module_id	fetch_manager_team_insights (as module_id in a learner drill-down)	fetch_manager_team_insights (as object_id, with object_type set to module)
Learner learner_id	fetch_manager_team_insights (as learner_id in a drill-down list, or as id in the Readiness Index lists)	fetch_manager_team_insights (as object_id, with object_type set to learner)

Note

A room has two IDs, and they aren’t interchangeable. fetch_room takes the UUID (room_id). edit_room and share_room take the numeric ID (room_id_int). Every tool that returns one returns both, so your AI client never needs to convert between them.

Choosing the right tool

Several tools can touch the same topic. This table shows which one fits which question.

If you want to	The AI client uses	What you get
Find content to read or add to a room	mindtickle_search	A ranked list of content items with their metadata and IDs, not a written answer.
Get a written answer about a deal, a call, a competitor, or your content	mindtickle_seller_copilot_ask	A generated answer with source links, based on content you can access, including Call AI recordings.
Find specific call recordings, or read a transcript or exact wording	find_call, then fetch_call	The matching calls, then each call’s summary, key moments, action items, or transcript.
See how a specific deal is going	find_deals, then fetch_deal_guide	The deal’s pains, objections, buying committee, competitors, and recommended next steps, if a Deal Guide exists for the deal.
Find a deal, company, contact, or room	find_deals, find_companies, find_contacts, find_rooms	Matching records, with the IDs other tools need.
Add a new buyer to a room’s guest list	create_contacts for an email address, or find_contacts for a person you can name, then create_room or edit_room	A contact_id, which is the only way to put someone on a guest list. share_room alone can’t do this.
See your own training progress, certificates, or badges	fetch_learner_insights	Your series, modules, scores, certificates, and badges. Seller Copilot doesn’t answer training status questions.
See how your team is doing, or find who is behind	fetch_manager_team_insights	Team readiness, training, coaching, and conversations, or a list of the learners behind on a series, module, or training category.
Find or open your AI Tutor tutorials (beta)	find_tutorial, then fetch_tutorial	A list of your tutorials with your progress, then the tutorial open in a screen inside the AI client.

Seller Copilot can also answer questions about your calls. Use the call tools when you need the recording itself, the transcript, or an exact quote. The call tools search recordings, not documents. For a documented win or loss analysis, a battle card, or a playbook, ask for that document by name.

Single-tool use cases

These examples each map to a single tool call. Your AI client selects the tool based on your prompt, so you don’t need to name the tool yourself. For the parameters each tool takes, see Supported tools.

Ask your AI client questions in plain language, the same way you’d ask a colleague.

Deals, companies, and contacts

What you want to do	Example prompt	Tool
Find a deal	“Pull up the Example Organization deal.”	find_deals
See which deals still need a room	“Which Example Organization deals don’t have a room yet?”	find_deals
See which deals already have a room	“Which Example Organization deals already have a room?”	find_deals
Find deals without a room in one stage	“Which deals in Negotiation don’t have a room yet?”	find_deals
Filter deals by stage	“Show all deals in the Negotiation stage.”	find_deals
Count deals	“How many deals are in Prospecting?”	find_deals
Check a deal’s value, close date, or owner	“What’s the value and close date of the Example Organization renewal, and who owns it?”	find_deals
Check a deal’s next step	“What’s the next step on the Example Organization renewal?”	find_deals
Find a company	“Do we have an account for Example Organization?”	find_companies
Page through matches	“Show me the next 10 companies matching Example Organization.”	find_companies
Find a contact by name	“Find Dana A. at Example Organization.”	find_contacts
Find a contact by email address	“Look up the contact with email dana@example.com.”	find_contacts
Check whether a contact agreed to be contacted	“Has dana@example.com agreed to be contacted?”	find_contacts
Add a buyer’s email address as a contact	“Add dana@example.com as a contact.”	create_contacts
Add several buyers as contacts at once	“Add dana@example.com, quinn@example.com, and kiran@example.com as contacts.”	create_contacts

Note

Stage names are free text from your CRM and must match exactly, so use the stage name as it appears in your CRM. Contact names match whole words, so “Dana” finds Dana but “Da” finds nothing.

create_contacts adds contacts only. It doesn’t put anyone on a room’s guest list and it doesn’t send email. If the address already belongs to a contact, you get that contact back and no duplicate is made. A new contact has no name or company, so a search by name won’t find it. Search by email address instead. Nothing in the Mindtickle MCP server edits or deletes a contact, and a mistyped address becomes a real contact record, so give the exact address. If you name a person instead of giving an address, ask your AI client to look them up first.

Digital Sales Rooms

What you want to do	Example prompt	Tool
Find a room	“Find the Example Organization room.”	find_rooms
Check whether a room exists for a deal or company	“Is there already a room for the Example Organization deal?”	find_rooms or find_deals
Rank rooms by buyer activity	“Show the most-visited active deal rooms.”	find_rooms
Find rooms with recent activity	“Which rooms had buyer activity most recently?”	find_rooms
Find draft rooms	“Show draft rooms.”	find_rooms
Find inactive rooms	“Show inactive rooms.”	find_rooms
Filter rooms by type	“Show all company rooms.”	find_rooms
Check a room’s owner and visits	“Who owns the Example Organization room, and how many people have opened it?”	find_rooms

Note

find_rooms returns every room you’re allowed to see, including colleagues’ rooms if you’re a manager or admin. Your AI client reads the owner name before it describes a room as yours. By default it returns only active rooms, so ask for draft or inactive rooms explicitly.

Content

What you want to do	Example prompt	Tool
Find a battle card	“Find the latest battle card for our top competitor.”	mindtickle_search
Find training modules	“Find training modules on discovery questions.”	mindtickle_search
Find sales assets	“Find pricing sheets and security FAQs.”	mindtickle_search
Find an asset hub	“Find the asset hubs about competitive intelligence.”	mindtickle_search

Seller Copilot

What you want to do	Example prompt	Tool
Prepare for a meeting	“Prep me for my call with Example Organization tomorrow.”	mindtickle_seller_copilot_ask
Catch up on an account	“Catch me up on the Example Organization account.”	mindtickle_seller_copilot_ask
Handle an objection	“How do I handle a pricing objection from Example Organization?”	mindtickle_seller_copilot_ask
Plan against a competitor	“What’s our best angle against our top competitor?”	mindtickle_seller_copilot_ask
Re-engage a stalled deal	“How do I re-engage a stalled deal?”	mindtickle_seller_copilot_ask
Ask a product or pricing question	“What are our current pricing packages?”	mindtickle_seller_copilot_ask
Ask about a named asset	“What does the Q3 pricing deck say about implementation timelines?”	mindtickle_seller_copilot_ask
Review recent calls	“What objections came up on my last 2 calls?”	mindtickle_seller_copilot_ask
Draft a follow-up	“Draft a follow-up email from my last call with Example Organization.”	mindtickle_seller_copilot_ask
Explore what Copilot can do	“What can you help me with?”	mindtickle_seller_copilot_ask

Note

Seller Copilot requires Mindtickle AI to be enabled for your organization and Copilot access granted to you by your admin. It answers from content you can already access. It doesn’t answer questions about your training status, such as assignments or completion, and it can’t summarize a module from only its name or link.

Calls

What you want to do	Example prompt	Tool
Find calls with an account in a time window	“Find my calls with Example Organization from last week.”	find_call
Find your most recent calls	“Show my 5 most recent calls.”	find_call
Find calls by topic	“Find calls where pricing came up.”	find_call
Find calls by topic in a time window	“Show calls about renewal pricing from the last 30 days.”	find_call
Find calls with a person	“What calls did I have with Sasha R. in the last 3 months?”	find_call
Find a call on a specific date	“Find my call with Example Organization on September 15.”	find_call

Note

find_call returns a list of matching calls, each with its participants, account, duration, and summary. It doesn’t return transcripts, key moments, or action items. For those, see Multi-tool use cases. The Deal Guide tool needs an ID from find_deals, so every Deal Guide example is in that section too.

Your own training

What you want to do	Example prompt	Tool
See your overall progress	“What’s my learning progress?”	fetch_learner_insights
List the modules you haven’t finished	“What do I still need to finish?”	fetch_learner_insights
List your completed series	“Which of my series have I completed?”	fetch_learner_insights
Find your certificates	“Do I have any certificates I can share?”	fetch_learner_insights
See your badges	“Which badges have I earned?”	fetch_learner_insights
See your oldest completed modules	“What did I finish first?”	fetch_learner_insights

Note

fetch_learner_insights returns only your own training data and needs no scope. Use it for questions about your training status. Seller Copilot doesn’t answer those.

Manager Insights

What you want to do	Example prompt	Tool
See a snapshot of your team	“How is my team doing?”	fetch_manager_team_insights
List the learners behind on a training category	“Who is overdue on training?”	fetch_manager_team_insights
Compare your team’s progress across series	“Which of my series is my team furthest behind on?”	fetch_manager_team_insights
See the lowest Readiness Index scores	“Who are my weakest reps on readiness compared to last quarter?”	fetch_manager_team_insights
Count your own coaching activity	“How many coaching sessions have I done this week?”	fetch_manager_team_insights
Find calls waiting for your review	“Which of my team’s calls should I review?”	fetch_manager_team_insights

Note

Manager Insights covers your direct reports. Coaching counts are your own reviews and sessions, not your reps’. This tool only reads data. It can’t assign training or send messages. To go from a series or a category to a person, see Track your team with Manager Insights.

AI Tutor (beta)

What you want to do	Example prompt	Tool
List your tutorials with your progress	“What tutorials do I have?”	find_tutorial
Find a tutorial by name	“Find my tutorial on objection handling.”	find_tutorial

Note

The AI Tutor tools are in beta and work only for companies that have AI Tutor turned on. Opening a tutorial takes two tools. See Learn and practice.

Multi-tool use cases

Some tasks need more than one tool call in sequence, with your AI client deciding which tool to call next based on the result of the previous one. The walkthroughs below are grouped by what you’re trying to do. Each one lists the example prompt, the steps your AI client takes, and the tools involved.

Create rooms

Create a room for a deal

Use this when a deal needs a room and you want content attached from the start.

Example prompt: “Build a room for the Example Organization renewal with the latest pricing sheet and security FAQ.”

1. The AI client calls find_deals with the deal name. Each result includes the deal’s deal_id, its company_id, and the room IDs if a room is already linked to the deal.
2. If a room already exists, the AI client tells you instead of creating a second one. A draft room counts as no room in the with_room filter, so the AI client checks the room IDs in the result before it creates anything.
3. The AI client calls mindtickle_search with fetchAssets set to true to find the pricing sheet and the security FAQ, and reads each asset’s id.
4. The AI client confirms the room name, type, deal, and number of assets with you, then calls create_room with room_type: deal, the deal_id, and the asset IDs in asset_ids. create_room can’t create an empty room, so the assets go in with this call.
5. The room goes live immediately, but nobody is emailed. The AI client gives you both links: the buyer link and your own internal link. It also compares total_items in the result with the number of assets it sent, to confirm they all attached.

This sequence uses find_deals, mindtickle_search, and create_room.

Create a room for an account

Use this for a room that belongs to a company rather than a single deal.

Example prompt: “Create a room for Example Organization with the security FAQ and the ROI calculator.”

1. The AI client calls find_companies to resolve Example Organization to a company_id.
2. The AI client calls find_rooms with the company name to check whether a room already exists. If one does, the AI client can add the assets to it instead. See Add content to a room.
3. The AI client calls mindtickle_search with fetchAssets set to true to find the assets, and reads each asset’s id.
4. The AI client calls create_room with room_type: company, the company_id, and the asset IDs in asset_ids.
5. The room goes live immediately, but nobody is emailed. To invite buyers, continue with Create a room and invite a buyer.

This sequence uses find_companies, find_rooms, mindtickle_search, and create_room.

Create a room and invite a buyer

Use this to create a room and email the invitation in one conversation.

Example prompt: “Create a room for Example Organization with the security FAQ and invite dana@example.com.”

1. The AI client calls find_companies to resolve Example Organization to a company_id. For a deal room, it calls find_deals instead and also reads the deal_id.
2. The AI client calls mindtickle_search with fetchAssets set to true to find the security FAQ, and reads its id.
3. The AI client calls find_contacts with the email address. If it finds a match, it reads the contact_id. If the contact’s consent is false, the AI client tells you and asks you to confirm before it invites them. If nothing matches, the AI client calls create_contacts with the address, which finds or creates a contact and returns its contact_id. Neither call sends an email.
4. The AI client calls create_room with the asset IDs in asset_ids and the contact ID in contact_ids. This adds the buyer to the guest list without emailing them.
5. The AI client reads the full list of addresses back to you. After you confirm, it calls share_room with the room’s room_id_int and the email addresses in invitees. share_room is the only tool that sends email, and the emails can’t be recalled.
6. The AI client compares the contacts listed in the share_room result with the addresses you asked for. An address that doesn’t appear in the result wasn’t emailed.

Why the guest list step matters: New rooms are invite-only by default. On an invite-only room, share_room refuses any address that isn’t on the guest list or covered by the room’s allowed domains. If the room is open to anyone with the link, or the address matches an allowed domain, share_room works without the guest list step and creates a contact for a new address automatically. The only exception is an address whose domain your organization has blocked. When share_room refuses an address, nobody is emailed and the error reads “this email may not access this room”. The same message appears for a blocked domain, so it doesn’t say which reason applies. Check the room with fetch_room, add the person with edit_room, and try again.

This sequence uses find_companies, mindtickle_search, find_contacts, create_contacts, create_room, and share_room.

Manage existing rooms

Add content to a room

Use this when a room exists and you want to add another asset.

Example prompt: “Add the new pricing sheet to the room I shared with Example Organization.”

1. The AI client calls find_rooms with the company or deal name and reads the room’s room_id_int.
2. The AI client calls mindtickle_search with fetchAssets set to true to find the pricing sheet, and reads its id and name.
3. The AI client calls edit_room with the room_id_int, add_assets set to true, and the asset in assets_to_add. The new content goes after what the room already holds. No email is sent, but buyers see the change at once.
4. The AI client checks the result before it reports success. If all_contents_already_exists is true, the asset was already in the room. If selected_resource_not_found is true, nothing was added.

Note

Content can only be added to a room. No tool removes or reorders it.

This sequence uses find_rooms, mindtickle_search, and edit_room.

Add a buyer to a room and invite them

Use this to give a new buyer access to a room that already exists.

Example prompt: “Add Kiran P. (kiran@example.com) to the Example Organization room and email them the link.”

1. The AI client calls find_rooms and reads the room’s room_id and room_id_int.
2. The AI client calls fetch_room with the room_id and include_guests set to true. This returns the current guest list, and whether the room is invite-only.
3. If Kiran is already a guest, the AI client skips to the last step. Otherwise, it calls find_contacts with their email address, or create_contacts if there’s no match, and reads their contact_id.
4. The AI client calls edit_room with the room_id_int and a contact_ids list that contains every current guest plus Kiran. edit_room replaces the whole guest list, so anyone left out is removed.
5. The AI client reads the address back to you. After you confirm, it calls share_room with the room_id_int and their email address.

Note

Adding a guest doesn’t email them. If you only want to add Kiran to the guest list, say so, and the AI client stops before share_room. If fetch_room shows that the room isn’t invite-only, or that their address matches one of the room’s allowed domains, the AI client can skip the guest list steps and go straight to share_room.

This sequence uses find_rooms, fetch_room, find_contacts, create_contacts, edit_room, and share_room.

Add several buyers to a room and invite them

Use this when you have a list of buyer email addresses, for example after a meeting with a buying committee.

Example prompt: “Add dana@example.com, quinn@example.com, and kiran@example.com to the Example Organization room and email them the link.”

1. The AI client calls find_rooms and reads the room’s room_id and room_id_int.
2. The AI client calls fetch_room with the room_id and include_guests set to true, to read the current guest list and whether the room is invite-only.
3. The AI client calls create_contacts once, with all the addresses in emails. For each address, it reads created and error. It reports any address that failed, and tells you which addresses are new contacts.
4. The AI client calls edit_room with the room_id_int and a contact_ids list that contains every current guest plus the new contact_id values. Addresses that failed in the previous step are left out.
5. The AI client reads the full list of addresses back to you. After you confirm, it calls share_room with the room_id_int and the addresses in invitees.
6. The AI client compares the contacts listed in the share_room result with the addresses you asked for. An address that doesn’t appear wasn’t emailed.

Note

Both tools limit how many addresses one call accepts, so your AI client splits a longer list across several calls. The find_contacts check is optional here. create_contacts returns the existing contact when an address is already known, so it never makes duplicates. Ask your AI client to look each person up with find_contacts first if you want to see their consent status before you invite them.

This sequence uses find_rooms, fetch_room, create_contacts, edit_room, and share_room.

Remove a buyer from a room

Use this to take someone off a room’s guest list.

Example prompt: “Remove Izumi T. from the Example Organization room’s guest list.”

1. The AI client calls find_rooms and reads the room’s room_id and room_id_int.
2. The AI client calls fetch_room with the room_id and include_guests set to true, and finds Izumi’s contact_id in the guest list.
3. The AI client calls edit_room with the room_id_int and a contact_ids list that contains every guest except Izumi.

Note

contact_ids needs at least one entry, so this can’t remove the last guest. Removing a guest changes the guest list only. If the room is open to anyone with the link, the person can still open it.

This sequence uses find_rooms, fetch_room, and edit_room.

Review a room’s content and engagement

Use this to see what’s in a room and how much buyers have used it.

Example prompt: “What’s in the Example Organization room, and how many people have opened it?”

1. The AI client calls find_rooms and reads the room’s room_id.
2. The AI client calls fetch_room with the room_id and include_content set to true.
3. The AI client reports the content items and the room’s totals: visitors_count (different people who opened the room), visits_count (total opens), room_contacts_count (invited guests), conversation_threads_count, and total_items.

Note

fetch_room returns engagement totals only. It doesn’t break engagement down by guest or by content item.

This sequence uses find_rooms and fetch_room.

Check a deal’s room from the deal

Use this when you think of the room as part of the deal. You don’t need to search for the room separately.

Example prompt: “Has anyone opened the room for the Example Organization renewal?”

1. The AI client calls find_deals with the deal name. The result includes the linked room’s room_id if the deal has one.
2. The AI client calls fetch_room with that room_id and reports visitors_count and visits_count.

This sequence uses find_deals and fetch_room.

Check who can open a room

Use this to review a room’s access settings before you share it.

Example prompt: “Who can open the Example Organization room, and does it expire?”

1. The AI client calls find_rooms and reads the room’s room_id.
2. The AI client calls fetch_room with the room_id and include_guests set to true.
3. The AI client reports the access settings: invited_contacts_only, requires_email, requires_passcode, requires_expiration_date with expires_at and expired, and domain_restricted. It also lists the guests.

Note

The passcode itself is never returned, only whether one is required. The allowed domains are returned only to someone who can edit the room. requires_email always reads true on an invite-only room.

This sequence uses find_rooms and fetch_room.

Change a room’s access or features

Use this to add a passcode or expiry date, restrict email domains, or switch buyer features on or off.

Example prompt: “Add a passcode to the Example Organization room and make it expire on December 31.”

1. The AI client calls find_rooms to get the room_id and room_id_int, then fetch_room to read the room’s current settings.
2. The AI client asks you for the exact passcode and confirms the expiry date with you. It doesn’t invent either value.
3. The AI client calls edit_room with the room_id_int, requires_passcode and passcode, and requires_expiration_date and expires_at. Only the fields it passes change.
4. The AI client calls fetch_room again to confirm the change took effect.

Note

edit_room can change more settings. See edit_room in Supported tools for the full list. A field you pass replaces its current value, so for a change such as “add example.com to the allowed domains” the AI client reads the current list first and sends the complete new list. Each entry in domains must start with @, for example @example.com. Changes reach buyers at once, and no email is sent.

This sequence uses find_rooms, fetch_room, and edit_room.

Rename a room or move it to another deal

Use this to retitle a room, update its welcome message, or attach it to a different deal or company.

Example prompt: “Rename the Example Organization room to Example Organization Proposal v2 and link it to the Example Organization renewal deal.”

1. The AI client calls find_rooms and reads the room’s room_id_int.
2. The AI client calls find_deals and reads the renewal’s deal_id.
3. The AI client calls edit_room with the room_id_int, the new name, and the deal_id.

Note

To change the welcome message, the AI client passes description. To attach the room to a different company, it passes company_id from find_companies.

This sequence uses find_rooms, find_deals, and edit_room.

Prepare for calls and follow up

Get a deal’s Deal Guide

Use this to see how a deal is going, or to prepare for a conversation about it.

Example prompt: “How is the Example Organization deal going?”

1. The AI client calls find_deals with the deal name and reads the deal’s deal_guide_id. If it’s null, no Deal Guide has been generated for that deal. The AI client tells you so and stops. Most deals don’t have one.
2. The AI client calls fetch_deal_guide with that deal_guide_id. The tool doesn’t accept a deal name, a deal ID, or a CRM opportunity ID.
3. The AI client checks not_found and processing_status before it reads any section. If processing_status is IN_PROGRESS, a guide is being built right now, and the AI client tells you to try again shortly.
4. The AI client reports the sections that answer your question. For a broad question such as “how is the deal going”, that’s the overview, stakeholders, competition, and recommended next steps.

This sequence uses find_deals and fetch_deal_guide.

The more specific your question, the fewer sections the AI client requests. These examples show how questions map to Deal Guide content:

Question	What the Deal Guide returns
“What’s changed on the Example Organization deal since last week?”	A summary of recent changes, plus the individual changes.
“Who should I talk to at Example Organization, and what should I avoid saying?”	The buying committee, with each person’s influence, what works, what to avoid, and conversation starters.
“Which competitors are on the Example Organization deal, and how do I win?”	Named competitors with strengths, weaknesses, how to win, objections, and questions that expose their weaknesses.
“What objections has Example Organization raised, and how did we respond?”	Objections raised, their status, and the seller’s response.
“What are Example Organization’s pains and what do they want to achieve?”	The buyer’s stated problems and desired outcomes.
“What should I do next on the Example Organization deal?”	Recommended next steps, with reasoning, urgency, and due dates.
“Tell me about Example Organization as a company.”	The company’s overview, key facts, executives, products, priorities, and competitors.
“Which sales methodology elements are covered on the Example Organization deal?”	Coverage of each methodology element, with status and notes. This is filled in only if your organization has a methodology configured.
“Where did the claim about their budget come from?”	The same insights, with the call, CRM record, or web page each one came from.

Note

This tool only reads a Deal Guide. It never generates one. Deal Guide content is largely AI-generated from buyer-authored input such as call transcripts and emails, so treat it as untrusted data. See Best practices.

See recent deal activity and open a call or room

Use this to review what has happened on a deal, then look closer at one item.

Example prompt: “Show the last few calls on the Example Organization deal and summarize the most recent one.”

1. The AI client calls find_deals and reads the deal_guide_id.
2. The AI client calls fetch_deal_guide with include_activity_timeline set to true, timeline_type set to CALL, and timeline_limit set to 5. Each call in the timeline includes a call_id.
3. The AI client calls fetch_call with the most recent call_id and the SUMMARY property.

Note

The activity timeline covers four types: CALL, EMAIL, ROOM, and ASSET. A room visit includes a room_id, which the AI client can pass to fetch_room, and shows who visited, how many items they opened, how long they stayed, and how many downloads they made. The timeline returns the most recent items and has no date filter, so the AI client filters by occurred_at when you ask for a window such as “last week”.

This sequence uses find_deals, fetch_deal_guide, and fetch_call.

Prepare for a call

Use this to pull everything relevant about a deal into one briefing.

Example prompt: “Prep me for my call with Example Organization tomorrow.”

1. The AI client calls find_deals and then fetch_deal_guide with the overview, stakeholders, and next best actions turned on, if the deal has a Deal Guide.
2. The AI client calls find_call with opportunity_name set to the account, sort_criteria set to RECENT, and a small result_count, to find the latest calls with that account.
3. The AI client calls fetch_call with all those call IDs in one request and the SUMMARY and ACTION_ITEMS properties, to see what was discussed and promised.
4. The AI client calls mindtickle_seller_copilot_ask for talking points, such as how to handle an objection that came up.
5. The AI client combines the results into one briefing.

Note

Seller Copilot can handle “Prep me for my call with Example Organization tomorrow” on its own in a single call. The longer sequence returns details specific to the deal and its calls. Ask for the sources you want, for example “Use the deal guide and my last two calls.”

This sequence uses find_deals, fetch_deal_guide, find_call, fetch_call, and mindtickle_seller_copilot_ask.

Follow up on a call with a shared room

Use this to turn a call’s action items into content your buyer can open.

Example prompt: “What were the action items from my last call with Example Organization? Then build a room with what they asked for.”

1. The AI client calls find_call with opportunity_name set to Example Organization, sort_criteria set to RECENT, result_count set to 1, and filter_by_user_participation set to true.
2. The AI client calls fetch_call with that call’s id and the ACTION_ITEMS property, and lists the follow-ups, such as sending a pitch deck or answering a security questionnaire.
3. The AI client calls mindtickle_search with fetchAssets set to true to find externally shareable assets that match, such as security FAQs, ROI calculators, competitive comparisons, and pricing sheets, and returns a shortlist.
4. You choose assets, and the AI client creates the room. See Create a room for an account, or Create a room and invite a buyer to email the invitation too.

This sequence uses find_call, fetch_call, mindtickle_search, and the room tools from the linked walkthroughs.

Summarize several calls

Use this to see how an account’s conversations have developed over a period.

Example prompt: “Summarize my calls with Example Organization from the last month.”

1. The AI client calls find_call with opportunity_name set to Example Organization, relative_time_value set to 1, relative_time_unit set to MONTH, and filter_by_user_participation set to true.
2. The AI client calls fetch_call once, with all the call IDs and the SUMMARY property.
3. The AI client writes one combined summary.

Note

Use a relative window such as “the last month” or an exact range such as “between February 1 and February 28”. A search can use one or the other, not both.

This sequence uses find_call and fetch_call.

Check what was said on calls

Use this when you need the actual words, not a summary.

Example prompt: “Did Example Organization mention our top competitor on any of our calls?”

1. The AI client calls find_call with opportunity_name set to Example Organization to find the calls.
2. The AI client calls fetch_call once, with all the call IDs and the TRANSCRIPT property.
3. The AI client searches the transcripts and tells you which calls mention the competitor, and what was said.

Note

SUMMARY, KEY_MOMENTS, and ACTION_ITEMS are condensed views of a whole call, and they often leave out detail. The AI client requests TRANSCRIPT when your question depends on what was said, who said it, or a specific part of the call, such as “How did my last call with Example Organization wrap up, and what was the final ask?” If a result’s truncated field is true, that value was cut short.

This sequence uses find_call and fetch_call.

Review calls on a topic across accounts

Use this to find patterns, such as which objections come up most.

Example prompt: “Which of my calls in the last 3 months had security objections? Give me the key moments.”

1. The AI client calls find_call with search_text set to the topic, relative_time_value set to 3, relative_time_unit set to MONTH, and filter_by_user_participation set to true.
2. The AI client calls fetch_call once, with all the call IDs and the KEY_MOMENTS property.

Note

The AI client puts a company or account name in opportunity_name and a person’s name in participant_names. It uses search_text only for what was discussed, never for a name.

This sequence uses find_call and fetch_call.

Research accounts and competitors

Review an account

Use this to see an account’s deals, contacts, and rooms in one answer.

Example prompt: “Give me a snapshot of Example Organization: its deals, contacts, and rooms.”

1. The AI client calls find_companies to resolve Example Organization to a company_id.
2. The AI client calls find_deals with company_ids, find_contacts with company_id, and find_rooms with the company name. These three calls don’t depend on each other, so the AI client can run them together.
3. The AI client combines the results into one summary.

This sequence uses find_companies, find_deals, find_contacts, and find_rooms.

List the contacts at an account

Use this before outreach to see who you can contact at an account.

Example prompt: “Who are our contacts at Example Organization, and have they agreed to be contacted?”

1. The AI client calls find_companies to resolve Example Organization to a company_id.
2. The AI client calls find_contacts with that company_id. Each contact includes a name, email address, job title, and consent status.

Note

consent is true if the person agreed to be contacted, false if they declined, and empty if there’s no record either way. share_room doesn’t enforce it, so check it before you invite someone.

This sequence uses find_companies and find_contacts.

Prepare to compete

Use this to get ready for a deal where a specific competitor is involved.

Example prompt: “Help me prepare to compete against our top competitor at Example Organization.”

1. The AI client calls find_deals and then fetch_deal_guide with the competition section turned on. It returns each competitor’s strengths, weaknesses, how to win, objections, and questions that expose their weaknesses.
2. The AI client calls mindtickle_search to find battle cards and competitive positioning guides for that competitor.
3. The AI client calls mindtickle_seller_copilot_ask for your best angle against that competitor.
4. If you ask whether Example Organization has mentioned that competitor, the AI client calls find_call and then fetch_call with the TRANSCRIPT property.

This sequence uses find_deals, fetch_deal_guide, mindtickle_search, mindtickle_seller_copilot_ask, and optionally find_call and fetch_call.

Track your team with Manager Insights

Find who on your team is behind on a series

Use this to see which reps haven’t finished a series.

Example prompt: “Who on my team hasn’t finished the onboarding series?”

1. The AI client calls fetch_manager_team_insights with object_type set to series and no object_id. The result lists your team’s series, each with a series_id, so the AI client can find the onboarding series by name.
2. The AI client calls fetch_manager_team_insights again with object_type set to series and that series_id as object_id. The result lists your learners in the series, most behind first.
3. If page.has_more is true, the AI client asks for the next page before it says who is behind.

This sequence uses fetch_manager_team_insights twice.

Check how one rep is doing

Use this to see one rep’s progress across everything assigned to them.

Example prompt: “How is Kiran P. doing on their training and missions?”

1. The AI client calls fetch_manager_team_insights with a drill-down that lists learners, such as object_type set to training and object_id set to overdue, and finds Kiran’s learner_id. Learners are identified by user ID and email address, not by name alone.
2. The AI client calls fetch_manager_team_insights with object_type set to learner and that learner_id as object_id.
3. If the result has no data, the learner isn’t one of your direct reports, and the AI client tells you so.

This sequence uses fetch_manager_team_insights twice.

Learn and practice

Plan what to study next

Use this to turn your unfinished modules into a short study list.

Example prompt: “What do I still need to finish, and what should I look at first?”

1. The AI client calls fetch_learner_insights with module_status set to NOT_STARTED and IN_PROGRESS, and the sections you don’t need turned off.
2. The AI client reads each module’s due date and, if has_more is true, asks for the next page before it ranks them.
3. If you ask for related reading on one module, the AI client calls mindtickle_search with that module’s topic.

This sequence uses fetch_learner_insights and, optionally, mindtickle_search.

Open a tutorial by name

Use this to open an AI Tutor tutorial when you know its name. The AI Tutor tools are in beta.

Example prompt: “Open my tutorial on objection handling.”

1. The AI client calls find_tutorial with search set to the tutorial name and reads the tutorial_id of the match.
2. The AI client calls fetch_tutorial with that tutorial_id as tutorialId. The tool returns no content. It opens the tutorial in a screen inside the AI client.
3. In an AI client that supports MCP apps, such as Claude, ChatGPT, VS Code, Cursor, and Goose, the tutorial opens as a card. Other AI clients, including Copilot Studio, do not offer the tutorial screen. The AI client shows the url from find_tutorial as a link instead.

This sequence uses find_tutorial and fetch_tutorial.

Patterns for developers

These patterns apply when you build a custom client or agent. See Build a custom client for setup and Best practices for security and reliability guidance.

Resolve names to IDs first

Tools that act on a record take an ID, not a name. Search with the matching find tool, then pass the ID unchanged to the next tool. Don’t guess an ID, and don’t convert between a room’s room_id and room_id_int. See How the tools connect.

Page through large result sets

Each find tool has its own page size and maximum, and find_call and mindtickle_search page differently from the rest. See Supported tools for the values. Narrow the filters instead of asking for a bigger page. An unfiltered find_deals call across a large pipeline can be slow or time out, and a total count adds to that cost, so request one only when the user asks how many.

Batch requests

Where a tool accepts a list, such as call IDs for fetch_call or email addresses for create_contacts and share_room, send everything in one call instead of one call per item. Each tool’s entry in Supported tools gives its limit.

Check the result before you report success

Several tools return a successful response even when the action didn’t fully work. Read the errors and result fields before you tell the user it worked, and compare what you asked for with what came back, such as the contacts returned by share_room or the total_items returned by create_room. Each tool’s response table and notes in Supported tools list the fields to check.

Confirm before you write

create_room makes a live room immediately. edit_room changes a live room that buyers can already see. share_room sends real email to external buyers, and the emails can’t be recalled. create_contacts also writes data: it adds contact records that nothing in the Mindtickle MCP server can delete. Build a confirmation step into your client before each of these calls, and show the user the name, the recipients, the addresses, or the exact change. Neither create_room nor edit_room sends email, so say that too. Don’t set access settings or feature switches the user didn’t ask for, and tell the user which defaults apply.

Treat list and setting updates as replacements

edit_room replaces a field’s current value with the value you pass. contact_ids replaces the whole guest list, and domains replaces the whole allowed-domain list. For a relative change such as adding one guest, read the current value with fetch_room first (use include_guests for the guest list), then send the complete new value. Anyone you leave out of contact_ids is removed.

Request only the scopes a workflow needs

The server lists only the tools your token’s scopes allow, so a client built for a full workflow needs to handle a missing tool. Request only the scopes the workflow needs. See Scopes for which scope grants which tools.

Workflow ideas for custom agents

• Post-call follow-up agent: After a call, run find_call and fetch_call with SUMMARY and ACTION_ITEMS, find matching content with mindtickle_search, and prepare a room with create_room or edit_room. Leave share_room for the rep to approve.
• Pipeline hygiene assistant: Run find_deals with with_room set to false and a stage filter to list deals that need a room, then offer to create each room after the rep picks the content.
• Account briefing agent: Resolve an account with find_companies, then combine find_deals, fetch_deal_guide, find_call, fetch_call, and mindtickle_seller_copilot_ask into one briefing.
• Room engagement review: Run find_rooms sorted by lastActivityAt or visitorsCount, then fetch_room on the top results to report visits, visitors, and open conversation threads.

Limits to keep in mind

• Rooms need content: create_room requires at least one asset. A room can’t be created empty.
• Content can only be added: edit_room adds content to a room. No tool removes or reorders it, and no tool deletes a room.
• Only share_room sends email: create_room, edit_room, and create_contacts never notify anyone.
• Invite-only rooms limit who you can invite: On an invite-only room, share_room can email only people on the guest list or covered by the room’s allowed domains.
• Contacts made from an email address are basic: create_contacts adds a contact with no name or company. Nothing in the Mindtickle MCP server edits or deletes a contact, and a mistyped address becomes a real contact record.
• Engagement is shown as totals: fetch_room returns visit and visitor counts, not a breakdown by guest or by content item.
• Room searches include colleagues’ rooms: find_rooms returns every room you can see, not only the ones you own.
• Contact names match whole words: A partial name such as “Da” finds nothing. Use the full name or the email address.
• CRM data can lag: Companies and contacts come from your CRM, so a very recent change may not appear yet.
• Not every deal has a Deal Guide: The tool reads an existing guide and never generates one. It has no date filter on its activity timeline.
• Call tools search recordings you can access: They don’t return calls you aren’t permitted to see, and they don’t search documents such as battle cards or win and loss reports.
• Seller Copilot has a defined scope: It doesn’t answer training status questions, and it can’t summarize a module from only its name or link.
• Training data is your own: fetch_learner_insights returns only your own data. It can’t return another learner’s progress.
• Manager Insights covers your direct reports: fetch_manager_team_insights returns no data for a learner outside your direct reports. Coaching counts are your own reviews and sessions, not your reps’. The tool can’t assign training or send messages.
• Screens need MCP app support: Tutorial cards show only in AI clients that support MCP apps. Other AI clients do not offer the tutorial screen. The AI client shows the url from find_tutorial as a link instead. The AI client can’t read what is inside a screen.
• The AI Tutor tools are in beta: They work only for companies that have AI Tutor turned on.

Tips for effective prompts

• Name the company, deal, or contact specifically. “Pull up the Example Organization deal” works better than “pull up my deal.”
• Name people, not roles. The tools look up contacts by name or email address, so “add Kiran P.” works and “add Dana A.’s manager” doesn’t.
• Type email addresses exactly. A typo creates a real contact record.
• Use stage names exactly as they appear in your CRM, such as “Negotiation”.
• For calls, say whose calls and when. “My calls with Example Organization last week” restricts the search to calls you were on. Give exact dates for a named range.
• Say what you need from a call. Ask for a “summary” or “action items” for a general view, and ask “what exactly did they say about pricing” when you need the words.
• Say whether to email. For rooms, tell your AI client when to stop short of inviting buyers, for example “create the room but don’t email anyone yet.”
• If your AI client asks a follow-up question, such as confirming a room title or a guest’s email, answer it directly. Your AI client’s own confirmation behavior determines when it asks.
• If a result doesn’t look right, ask again with more detail rather than assuming the tool returned nothing relevant. Vague prompts return vague matches.

Related

• Mindtickle MCP: What the server does and how to decide which connection path applies to you.
• Supported tools: The full list of tools, what each one does, and the scope it requires.
• Best practices: Security, reliability, and integration design.