Connect Five Dots.
Start with one question.
Bring the RevOps method into your AI assistant, connect the context you choose, and work through your first question together.
No account yet? Start free: the full audit, with no card and no time limit. Upgrade any time you like.
Three things to have ready.
Your Five Dots access.
Sign up for a free account, or accept the invitation from your team. Each person connects with their own access.
Your AI workspace.
You need your own AI subscription and permission to add a custom connection. Your provider’s plan and your workspace administrator determine what is available.
A question and some context.
Start with your business brief, existing decisions and relevant documents. Add your CRM or other tools when the question needs them.
Set up in your AI assistant.
Add Five Dots as a custom connection using the steps below. Availability depends on your assistant and workspace permissions.
Get your connection details.
Sign in to your Five Dots account and open the connection setup. The connector URL is:
https://app.fivedots.ai/api/mcp. Nothing else to enter; you sign in when you connect.Add it in Claude.
Claude: open Customize → Connectors → + Add → Add custom connector. Name it Five Dots, paste the URL above and continue with the detected OAuth settings.
Add it in ChatGPT.
ChatGPT: open Plugins → + → Add custom MCP server. Name it Five Dots, paste the URL above and select OAuth. Review the warning and choose Create as a plugin.
Sign in and enable the connection.
Follow the steps for your chosen assistant. Sign in with your Five Dots account. In Claude, enable the connector in your conversation. In ChatGPT, install the personal plugin, open a new Work chat and select @Five Dots.
Start a conversation.
Ask your assistant to use Five Dots. It should read the method, establish the scope with you and explain what context is needed.
“Use Five Dots to review our sales process. Start with our business brief and existing documents, and tell me what else you need.”
Menu names can change. Claude’s current connector guide ↗︎
Check the workspace,
as well as the assistant.
Connection options differ by product, plan and administrator settings. A provider’s support for custom connections does not confirm that every Five Dots workflow has been tested in your setup. We can help check yours before you start.
On managed workspaces, an owner or authorised administrator adds the connector first. Each member then connects with their own account.
Use the custom MCP server steps above. Account and workspace policies determine who can add it. OpenAI’s app setup guide ↗︎
Confirm which Copilot product you use. Microsoft documents an MCP connection route in Copilot Studio; that is a specific agent-building setup. Copilot Studio guide ↗︎
Tell us which Gemini product or workspace you use so we can check the connection route with you. Setup instructions depend on the environment.
Ask your workspace administrator which custom integrations are enabled, then contact us to confirm the Five Dots setup for your workspace.
Add Five Dots as a remote MCP server using your connector URL, then complete the sign-in prompt. Your team may need to allow the connection. Cursor’s MCP setup guide ↗︎
What is MCP?
MCP is the connection standard used to give an AI assistant access to tools. Five Dots uses it to supply its method. Your CRM and document connections are set up separately in your AI workspace.
Nine tools,
and what to say.
Five Dots supplies method and expertise through nine read-mostly tools. Your assistant chooses the tool; you talk to it normally. None of them reads your systems: that is your assistant, through the connectors you enable.
get_playbook- The session guide: how your request is understood, where your work is kept, how owners and decisions are recorded, and how the GTM Audit is planned, run and scored. Your assistant reads it first.
list_modulesstart_runget_instructions- The audit. What it covers, by pillar and module; a plan for the scope you choose, with a run id and a due date per module; and each module’s steps, scoring criteria and remediation playbook.
search_knowledgeget_knowledge- The knowledge base. RevOps and go-to-market frameworks, metrics, benchmarks and plays, found by topic and read in full, so a finding is interpreted against the method behind it.
list_skillsget_skill- The skills library. Executable go-to-market playbooks, carried out in your session against your own tools.
request_help- The help centre, in the chat. The articles on setup, data, roles and running the audit, plus an email draft to us for when they are not enough.
What to say.
- “What can I do with Five Dots?”
- “What does the GTM Audit cover, and how long does it take?”
- “Plan a GTM Audit for sales only, to be done by the end of next month.”
- “What is a healthy net revenue retention for a Series A SaaS company?”
- “Help me set up a cold-email programme with the Five Dots skills library.”
- “Continue the Acme Five Dots audit. My module is pipeline hygiene.”
From a terminal or an IDE.
The same connector URL works in command-line assistants. Each opens your browser to sign in.
# Claude Code
claude mcp add --transport http five-dots https://app.fivedots.ai/api/mcp
# then /mcp inside Claude Code to sign in
# OpenAI Codex (CLI, IDE and app)
codex mcp add five-dots --url https://app.fivedots.ai/api/mcp
# Cursor: Settings → MCP → add a remote server with the URL above
# ChatGPT: Plugins → + → Add custom MCP server → OAuth → Create as a plugin → Install
Four habits that
make Five Dots better.
The method is the same for everyone. What it can do for you depends on what your assistant can see, what it remembers and who is in the room.
Connect more than the CRM.
Every tool your assistant can reach makes Five Dots better: CRM, knowledge base, marketing and billing platforms, Slack, the project board. One connector gives a thin picture. The full stack runs the audit on evidence.
Let it build a company memory.
On the first run your assistant creates a “Your company — Five Dots” workspace in a store you already use and keeps findings, decisions and a journal there. Every session starts by reading it, so it gets smarter each time. Keep that one workspace; do not start fresh.
Bring your team in.
Invite the people who own each area: sales for pipeline, marketing for lead flow, RevOps for data, finance for ARR. Each picks up their module from their own assistant, and the shared workspace keeps everyone on one record.
Run it where your team already talks.
If your company uses Claude in Slack, add Five Dots to that channel. Assignments and findings land where the team already reads, and anyone can continue their module from there.
Connect only the tools you want your AI assistant to use, with the permissions appropriate to the work.
No live CRM connection? Start with an interview or a small, relevant export in your own AI workspace. Missing evidence should be marked as not assessed. Larger datasets may need a sample or summary.
Review proposed changes and test them in a sandbox before putting them into production. What can be read or changed depends on the connectors and permissions available in your assistant.
Five Dots supplies the method. Business records and audit findings are not returned to the methodology service. Your AI provider and connected tools process information under your agreements with them. Read about data and permissions →
Need a hand?
Email our team.
Need help before you can connect?
help@fivedots.aiI cannot add the connection.
Check that your AI workspace permits custom connectors or apps. Ask your administrator if the option is missing, and confirm that your Five Dots access is active. If you still cannot connect, email us with the assistant name and the error message.
Five Dots cannot see my CRM or documents.
The Five Dots connection supplies the method. Enable your CRM or document connector separately in the same AI workspace and check your access to the source. You can continue with an interview or a relevant export while a connection is unavailable.
How do I continue a previous session?
Use the same connected workspace where you saved your findings and decisions. Ask your assistant to “continue where we left off with Five Dots” and have it read the saved record before continuing.
What does Five Dots receive, and what does it store?
Only the request for method: which part of the audit you chose, the names of the tools you connected, whether your Five Dots workspace exists yet and, for a knowledge-base search, a few generic topic words. Never your records, findings, people’s names or your company’s name. For your account we keep its name and plan, members’ names, work emails and roles, and run records. The privacy policy spells it out →
“This account cannot run tools right now.”
The account has no active plan, or its plan has ended. An admin of the account can manage it at app.fivedots.ai. Connecting and listing the tools keep working meanwhile.
“Couldn’t reach the MCP server”, or sign-in keeps looping.
Check the URL is exactly https://app.fivedots.ai/api/mcp, remove the connector and add it again, and sign in when asked. On a company plan, an administrator may need to allow custom connectors or the sign-in domain auth.fivedots.ai. If it still fails, email us the assistant name, the time and the message shown.
The tools look outdated after an update.
Updates to Five Dots reach your assistant on its next session; nothing to reinstall. If a tool list looks stale, remove and re-add the connector so your assistant reads it fresh.
Can Five Dots run on a schedule?
Yes, where your assistant supports scheduled tasks. Ask it for the “ready-made schedules” help article: it includes prompts for a weekly pipeline check and a monthly audit refresh that run from your saved workspace. On paid plans, ask which standing watch belongs to a fix: it re-reads the audit’s figures on a cadence and keeps a dated log in your workspace. Five Dots itself runs nothing between sessions.
How do I contact support?
Email help@fivedots.ai with what you tried, what happened and the tools involved. Leave out customer records, business figures and credentials.