AI for Sales Ops: Run HubSpot, Slack and Linear From One Chat Agent
MCP Playground
MCP Playground
๐ TL;DR
- Three official MCP servers: HubSpot (
mcp.hubspot.com), Slack (mcp.slack.com/mcp), Linear (mcp.linear.app/sse) - Ask questions that span all three tools in a single prompt โ no switching tabs
- Set up in under 5 minutes with the Sales Pipeline template in MCP Agent Studio
- Token requirements: HubSpot private app token ยท Slack bot token (xoxb-โฆ) ยท Linear personal API key
Sales ops teams live in three tools: HubSpot for pipeline, Slack for deal conversations, and Linear for tracking feature requests that come up in calls. The context you need to close a deal is split across all three, and pulling it together manually is slow.
This guide shows you how to connect all three to a single AI agent using their official MCP servers โ so you can ask "Which deals over $20k have had no Slack activity this week, and do any have open Linear requests?" and get a real answer without switching tabs.
What You Can Do With This Agent
Pipeline health in one prompt
"Which deals over $10k haven't had activity in 14 days?" โ pulls directly from HubSpot deal data.
Cross-tool deal context
"Find Slack messages about Acme Corp and any open Linear tickets linked to their account."
Personalised outreach drafting
"Draft a re-engagement email for Jane at Acme using her HubSpot profile and the last Slack thread."
Pipeline updates
Move a deal to "Closed Won", log a call note, and post a win summary to a Slack channel โ all in one prompt.
Feature request tracking
Log a customer feature request from a Slack thread as a Linear issue, tagged by account size or deal stage.
Weekly sales summaries
"Summarise this week's new deals, closed deals, and any Slack threads that need a follow-up."
The Three Official MCP Servers
| Tool | MCP Server URL | Transport | Token |
|---|---|---|---|
| HubSpot | https://mcp.hubspot.com | HTTP | Private app access token |
| Slack | https://mcp.slack.com/mcp | HTTP | Bot token (xoxb-โฆ) |
| Linear | https://mcp.linear.app/sse | SSE | Personal API key |
Step 1 โ Get Your Three Tokens
HubSpot private app token
- In HubSpot go to Settings โ Integrations โ Private Apps โ Create a private app
- Give it a name and set scopes:
crm.objects.contacts.read,crm.objects.deals.read,crm.objects.companies.readโ add write scopes if you want the agent to update records - Click Create app and copy the access token
Slack bot token
- Go to api.slack.com/apps โ create a new app โ OAuth & Permissions โ Bot Token Scopes
- Add:
channels:read,channels:history,search:read, and optionallychat:writeif you want the agent to post messages - Install to your workspace and copy the Bot User OAuth Token (
xoxb-โฆ) - Invite the bot to any channels it should read:
/invite @your-bot-name
โ ๏ธ Slack tier note: The search:read scope (used for "find Slack messages about X" prompts) requires a paid Slack plan โ Pro, Business+, or Enterprise. On the Free workspace tier, prompts that search across messages will fail. Lookup-by-channel prompts using channels:history still work on Free.
Linear personal API key
- In Linear go to Settings โ API โ Personal API keys โ Create key and copy it
Step 2 โ Connect in Agent Studio
Fastest path: use the pre-built template
The Sales Pipeline template has HubSpot, Slack and Linear pre-configured โ URLs, transports and system prompt included. Click it, paste your three tokens, and start chatting.
Use Sales Pipeline Template โTo set up manually in MCP Agent Studio:
- Click + Add Server โ
https://mcp.hubspot.com, select HTTP, paste HubSpot token - Add second server โ
https://mcp.slack.com/mcp, select HTTP, paste Slack bot token - Add third server โ
https://mcp.linear.app/sse, select SSE, paste Linear API key - Pick a model โ Claude Sonnet 4.6 for complex cross-tool chains; DeepSeek V3.2 or Gemini 3 Flash for routine daily lookups
- The agent auto-discovers all tools from all three servers on first message
10 Prompts to Start With
- "Which deals in my pipeline are at risk this month?"
- "Draft a personalised intro email to Jane at Acme using her HubSpot profile and company industry."
- "Find Slack messages mentioning 'pricing' from the #sales channel in the last 7 days."
- "Move deal #12345 to Closed Won and log a call summary."
- "List new contacts added this week grouped by industry."
- "Find any Linear tickets tagged as customer requests that came from deals in the negotiation stage."
- "Summarise the last Slack thread about Acme Corp and tell me the next action."
- "Create a Linear ticket for the dashboard feature Acme requested and link it to their deal."
- "Which deals close this quarter and haven't had contact in over 30 days?"
- "Post a Slack message to #sales-wins summarising today's closed deals from HubSpot."
Which Model to Use
- Claude Sonnet 4.6 โ best for complex cross-tool queries that need to join data across HubSpot + Slack + Linear in one answer, or draft context-aware emails. Worth the extra credit cost for these.
- DeepSeek V3.2 / V4 Pro โ solid for structured lookups (deal lists, contact filters). Significantly cheaper per run, good choice for routine daily queries once you know the pattern works.
- Gemini 3 Flash โ fast and cheap for simple single-server queries. Less reliable for three-server cross-tool chains; use it for "list my open deals" but not "find Slack context about this deal and draft a follow-up."
Run the same prompt across models with Compare Mode to find where the quality difference justifies the cost for your specific data.
Without the agent
- Open HubSpot, filter deals manually
- Switch to Slack, search for company name
- Switch to Linear, search for linked tickets
- Copy context into email draft manually
- Repeat for every deal
With the agent
- One prompt in chat
- Agent queries all three tools
- Single consolidated answer
- Draft, update or create in same thread
- Save as named agent to reuse daily
Set up your Sales Ops agent in under 5 minutes
Official HubSpot ยท Slack ยท Linear MCP servers ยท Any AI model ยท Free credits on sign-up
Frequently Asked Questions
Related Guides
Written by MCP Playground
15+ years in product development. AI enthusiast building developer tools that make complex technologies accessible to everyone.
Related Resources
Test any MCP server with 30+ AI models โ free
Connect any MCP endpoint and chat with Claude, GPT-5, Gemini, DeepSeek and more. Watch every tool call live.
โฆ Free credits on sign-up ยท no credit card needed