Say Goodbye to Copy-Paste: Your CRM Now Speaks Directly to Claude

What does "CRM speaks directly to Claude" actually mean?

Before MCP (Model Context Protocol), using an AI assistant with your CRM meant a laborious cycle: open Follow Up Boss, find a contact, copy their name and notes, paste into Claude, read the response, copy that back, paste it into a note or email field in FUB. Repeat dozens of times a day.

MCP changes the direction of that relationship. Instead of you ferrying data between two systems, Claude connects directly to Follow Up Boss through a secure, authenticated URL. When you ask Claude "who are my hottest leads right now?" it doesn't need you to paste anything — it calls your CRM, fetches the data, and reasons over it in a single turn.

Follow Up Ace implements this as a production-grade MCP server. The Claude endpoint is https://followupace.com/mcp (Streamable HTTP). The ChatGPT-compatible SSE endpoint is https://followupace.com/api/mcp/sse/. Both are authenticated to your account, so Claude only ever sees your data.

What can Claude actually do inside Follow Up Boss?

The connector exposes 200+ tools spanning the full FUB data model. Here is a representative set of what Claude can read, analyze, and act on in a single conversation:

• Contact operations — fetch people, search by name or tag, create new contacts, edit fields, check for duplicates, merge records.
• Pipeline management — fetch deals, move stages, create or edit deal records, review pipeline health with the pipeline-health-check prompt tool.
• Notes and tasks — create, fetch, edit, and delete notes; create tasks with due dates; mark tasks complete.
• Appointments — schedule, edit, and fetch upcoming appointments for any contact.
• Calls and texts — log calls, fetch call history, view inbound text threads.
• Action plans — apply action plans, update enrollment status, enroll contacts in automations.
• AI-generated messages — generate a draft email or SMS for a contact using the generate_message tool, tuned to their history.
• Lead nurture — surface stale leads using the lead-nurture-optimizer prompt, filtered by days since last contact and pipeline stage.
• Bulk operations — bulk-tag people, bulk-assign agents, bulk-apply action plans across a filtered contact set.
• Speed-to-lead — check Zillow lead response time with zillow_speed_to_lead_check and identify leads at risk of going cold.

All of this happens in one Claude conversation. You ask a question, Claude calls the right tool, and the answer comes back contextualized — no switching tabs, no manual search.

How do I connect Follow Up Boss to Claude?

Setup is a two-minute process. You need an active Follow Up Ace Pro plan (the MCP connector is a Pro feature — voiceChatEnabled and MCP access are both gated at the Pro tier in the platform's plan logic).

1. Sign up or sign in at followupace.com and confirm your plan is Pro.
2. Open Claude Desktop (or any MCP-compatible client) and navigate to its MCP settings or connector configuration.
3. Add a new MCP server with the URL: https://followupace.com/mcp
4. Authenticate — the connector uses your Follow Up Ace credentials. Follow the OAuth prompt or paste the API token shown in your Ace dashboard.
5. Test the connection — ask Claude "Fetch my last 5 contacts in Follow Up Boss." If it returns names from your CRM, the link is live.

For ChatGPT users, use the SSE endpoint instead: https://followupace.com/api/mcp/sse/. The same 200+ tools are available on both endpoints.

Real workflows agents run through Claude every day

Understanding the tools is one thing. Seeing how they chain together in a real workday is more useful. Here are four concrete workflows that replace a lot of manual clicking:

Morning pipeline review

Instead of opening FUB and scrolling through each pipeline column, ask Claude: "Run a pipeline health check and flag any contacts stuck in the same stage for more than 21 days." Claude calls the pipeline-health-check prompt tool, fetches your deals, and returns a prioritized list of contacts who need attention — with recommended next actions.

Lead nurture triage

Ask: "Find leads I haven't contacted in 30 days who are still in the Active stage, and draft a re-engagement text for each." Claude runs the lead-nurture-optimizer, fetches the matching contacts, and calls generate_message for each — outputting ready-to-send messages you can review before sending.

Contact deep-dive before a showing

Five minutes before meeting a buyer, ask: "Give me a full brief on Sarah Johnson — her notes, recent calls, current pipeline stage, and any open tasks." Claude fetches everything through get_all_person_data and fetch_notes, returning a clean summary. No hunting through tabs.

Zillow speed-to-lead check

Zillow leads go cold fast. Ask Claude: "Check my Zillow speed-to-lead status for this week — which new leads waited more than five minutes for a first contact?" Claude calls zillow_speed_to_lead_check and surfaces the at-risk contacts so you can course-correct. See the full playbook at followupace.com/zillow-playbook.

How does this compare to just asking Claude with pasted data?

Here is a direct comparison of the two approaches for a common task — drafting a follow-up email for a contact:

The difference compounds fast. An agent handling 20 contacts a day eliminates hundreds of context switches per week.

Is it safe to give Claude access to my CRM?

The MCP connector authenticates through your Follow Up Ace account credentials, so Claude only accesses the data your account is authorized to see — nothing more. The connection is scoped to your Follow Up Boss organization. No data is stored permanently in Claude; it reads from FUB in real time for each query.

Write operations (creating notes, editing contacts, creating tasks) require the same permissions you have inside FUB. Claude cannot do anything your account could not do manually.

One additional layer: Follow Up Ace includes a built-in compliance scanner that checks generated messages against Fair Housing and licensing rules before they reach your clients. The scanForComplianceViolations() function in the platform's compliance layer catches discriminatory language patterns automatically. See the compliance overview for details on what is checked.

What plan do I need, and what does it cost?

Follow Up Ace has three per-seat plan tiers — free, Regular ($25/seat/month), and Pro ($55/seat/month). The MCP connector, voice chat, and agentic tools require Pro. The free and Regular tiers include the embedded AI chat, ACE lead scoring, and contact analysis inside Follow Up Boss.

Teams with larger contact databases also have access to the Ace Trove — an account-wide intelligence layer priced by total contacts:

Ace Trove adds nightly account-wide contact analysis, ACE lead scoring fields written directly to Follow Up Boss custom fields, and the Seller Radar intelligence layer on top of the per-seat plan. See the Ace Trove page for the full feature breakdown.

What are the ACE scoring fields that Claude can read?

When Ace Trove is active, Follow Up Ace writes a set of custom intelligence fields to each contact in Follow Up Boss. Claude can fetch and reason over all of them. The current non-deprecated scoring fields include:

• ace_score — overall lead score for prioritization
• ace_tier — tier classification (e.g., hot / warm / cold)
• ace_status — current contact status derived from CRM activity
• ace_response_time — measured first-response time for inbound leads
• ace_velocity_score — engagement velocity based on recent activity patterns
• ace_days_since_inbound — days elapsed since the lead first arrived
• ace_preferred_channel — detected preferred communication channel
• ace_lead_type — buyer, seller, or renter classification
• ace_property_profile — property type and criteria from lead activity
• ace_search_area — geographic area of interest
• ace_buyer_readiness — readiness signal derived from engagement
• ace_lead_summary — AI-generated plain-English summary of the contact
• ace_next_action — recommended next step for this contact

Because these fields live in Follow Up Boss as standard custom fields, Claude can filter, sort, and analyze them the same way it would any other CRM data. Ask "show me all contacts with an ace_tier of hot and no task in the last 7 days" and Claude fetches the matching records and surfaces them directly.

Can I use ChatGPT instead of Claude?

Yes. Follow Up Ace publishes a separate SSE endpoint for ChatGPT and other OpenAI-compatible clients: https://followupace.com/api/mcp/sse/. The same 200+ tools are available. The setup process differs slightly by client — ChatGPT's plugin or connector settings will prompt for the SSE URL.

Claude and ChatGPT both connect to the same backend, so the data returned is identical. The difference is in how each AI reasons over it — Claude's extended context window and tool-use behavior can handle longer contact histories in a single turn, which is useful for pipeline reviews across large databases.

What does the agentic layer add beyond basic MCP?

The MCP connector gives Claude read/write access to FUB data. The agentic layer adds a coordination tier on top of that: multi-step prompt chains that combine several tool calls into a single workflow, Ace Trove integration for scoring and analysis, and context-aware briefing tools like daily-agent-briefing and contact-deep-dive.

In practice, the agentic tools let Claude run what would otherwise be a four-step manual workflow — find contacts, filter by scoring fields, generate messages, log activity — as a single conversational request. The tools handle the sequencing; you handle the review.

If you are evaluating how this stacks up against other AI-for-FUB options, the compare page breaks down the key differences by feature category. The guides section has step-by-step walkthroughs for common workflows.

Getting started: the fastest path to zero copy-paste

The shortest path from your current workflow to a Claude-connected CRM:

1. Start with a free Follow Up Ace account to confirm your Follow Up Boss connection works.
2. Upgrade to Pro when you want MCP access and voice features.
3. Add https://followupace.com/mcp to Claude Desktop or your MCP client of choice.
4. Run a morning pipeline review to see your first zero-copy-paste workflow in action.
5. Add Ace Trove if you want nightly contact scoring written back to FUB custom fields at scale.

The copy-paste habit formed because there was no better option. Now there is — your CRM and your AI assistant share the same data in real time, and the friction between "thinking about a contact" and "acting on a contact" disappears.

More: the AI real estate assistant for Follow Up Boss.