Connect Perplexity and AI Search to Your CRM Data

You are in a meeting and someone mentions a company you have never heard of. You know what they do, “that fintech startup that does invoice automation”, but not their exact name. You open your CRM, type what you think the name might be, and get zero results. You try Google, find the company, copy the name back into the CRM, create the record, and add the contact. Five minutes gone, and you are still not sure the company record is complete.

This is one of the most common friction points in CRM usage: search that only works when you already know what you are looking for. Booked55 solves this by integrating Perplexity AI search directly into its MCP server through the search_company_ai tool (giving your AI assistant the ability to find companies by description, not just by name.

The Search Problem in Traditional CRMs

CRM search is designed for exact or partial name matching. You type “Acme” and get Acme Corp, Acme Industries, and Acme Consulting. This works when you know the name. It fails when you:

• Heard a company referenced by its product, not its name
• Know the industry but not which specific company your contact works at
• Have a phonetic spelling from a voice note or business card
• Need to find a company by its former name after a rebrand
• Want to research competitors in a space before adding them to your CRM

These are not edge cases, they are daily occurrences for sales reps, recruiters, and consultants who constantly encounter new companies. Standard search forces you out of the CRM and into a browser, breaking the workflow and introducing data entry errors when you copy information back manually.

How search_company_ai Works

Booked55's MCP server includes two company search tools:

• search_company (Standard regex-compatible name search against your CRM database. Fast, precise, works for known names.
• search_company_ai) Perplexity-powered AI search that finds companies by description, product, industry, or alternative names. Use when standard search returns no results.

When your AI assistant (ChatGPT, Claude, or Booked55's built-in assistant) receives a company lookup request, it tries search_company first. If that returns nothing, it automatically falls back to search_company_ai, which queries Perplexity's AI search engine to identify the company and return structured results you can add to your CRM.

Practical Examples

In each case, the AI chains search_company_ai with create_company and optionally create_contact (turning research and data entry into a single conversational request.

MCP: The Bridge Between AI Search and CRM Data

Perplexity is an AI search engine. Your CRM is a structured database. Without MCP, connecting them requires custom code that takes search results and maps them to CRM fields. With MCP, the connection is built in:

1. You ask your AI assistant a research question about a company.
2. The AI calls search_company_ai on Booked55's MCP server.
3. The MCP server queries Perplexity, returns structured company data.
4. The AI presents the results and offers to create a CRM record.
5. You confirm, and the AI calls create_company with the enriched data.

This is the power of MCP as an integration layer: AI search capabilities and CRM operations coexist in the same tool ecosystem, orchestrated by the AI model based on your natural language intent. You do not switch between Perplexity and your CRM) they work together through the protocol.

AI Search Beyond Company Lookup

The Perplexity integration is one example of a broader pattern: AI-native CRMs embedding external intelligence into MCP tools. Other applications include:

• Contact enrichment (Booked55 auto-enriches contacts with LinkedIn, job history, and company data without a separate search step.
• Pre-meeting research) Ask your AI to pull CRM records AND search for recent news about a company before a call.
• Competitive intelligence, “Find the top three competitors to Acme Corp in the CRM analytics space” combines AI search with CRM context.
• Event prospecting, After a conference, use AI search to identify companies from attendee lists and add them to your CRM in batch.

As MCP servers evolve, expect more AI search integrations (market data, social signals, news monitoring) all accessible through the same conversational interface.

Connect Perplexity to Booked55 via MCP

Perplexity works as an MCP client, so you can connect it directly to Booked55 and manage CRM data from Perplexity conversations (not just use Perplexity search inside Booked55 tools.

1. Open Perplexity → Settings → Connectors.
2. Click + Custom connector → Select Remote.
3. Enter https://mcp.booked55.com
4. Complete OAuth authorization with your Booked55 account.
5. Available on Pro, Max, and Enterprise plans.

Once connected, ask Perplexity to search your CRM, create contacts, or research companies. It can chain search_company_ai with create and update tools in a single conversation.

Using Perplexity-Powered Search via MCP Clients

The search_company_ai tool is available through any MCP client connected to Booked55:

• ChatGPT, Connect via the Booked55 MCP app (see our setup guide).
• Claude Desktop, Add Booked55 as a custom connector or via mcp-remote (see our Claude guide).
• Booked55 AI assistant, Available in-app without any setup.

The search behavior is identical across clients, the AI selects the right tool based on your question, regardless of which assistant you are talking to.

Best Practices

• Search CRM first, Always check if the company already exists with search_company before creating a duplicate.
• Be descriptive, The more context you give (“the fintech company that automates invoice processing”), the better AI search results.
• Confirm before creating (Review the company the AI found before adding it to your CRM, especially for common names.
• Chain with contacts) After finding a company, immediately add the contact you met: “Create the company and add John Smith as VP Sales.”

Perplexity vs. Standard Web Search in CRM Context

You might wonder why Booked55 integrates Perplexity rather than relying on the AI model's built-in knowledge. Three reasons:

• Freshness (LLM training data has a cutoff date. Perplexity searches the live web, so company information includes recent funding rounds, rebrandings, and leadership changes.
• Structured results) The MCP tool returns company data in a format the AI can directly map to CRM fields, rather than unstructured text you would need to parse manually.
• Source attribution (Perplexity cites its sources, so you can verify company details before adding them to your CRM.

This is a concrete example of why MCP matters: it lets CRM vendors integrate best-in-class AI services (Perplexity for search, enrichment providers for contacts) into a unified tool layer that any AI client can orchestrate.

The Bigger Picture

Connecting AI search to CRM data through MCP is a preview of where business software is heading. Instead of siloed tools, a search engine here, a CRM there, a spreadsheet for notes (MCP creates a unified layer where AI orchestrates across systems based on what you need. Booked55's Perplexity integration is one tool in a growing MCP server. For the full tool reference, see the Available Tools documentation.

Real-World Example: Freelance Recruiter Prospecting

Kevin is a solo tech recruiter who hears company names on podcasts and at meetups) often imprecisely. Last week someone mentioned “that observability startup that competes with Datadog, based in San Francisco, raised Series B recently.” Standard CRM search returned nothing. In Claude Desktop he asked:

“Use AI search to find the observability startup in San Francisco that raised Series B and competes with Datadog. Create the company in my CRM and add a task to research their engineering leadership team.”

Claude called search_company_ai, identified the company, confirmed with Kevin, then chained create_company and create_task. Total time: under three minutes. Without MCP, Kevin would have opened Perplexity in a browser, copied results, switched to CRM, and manually created records (if he bothered at all. See what MCP is and AI contact enrichment for related workflows.

Technical Flow: From Question to CRM Record

Here is the exact sequence when an MCP client (Perplexity, Claude, or ChatGPT) finds and adds a company by description:

1. ChatGPT sends tools/call for search_company with your best-guess name) returns empty.
2. ChatGPT automatically retries with search_company_ai, passing your descriptive query to Perplexity via Booked55's backend.
3. Perplexity returns company name, domain, and context. The MCP server normalizes this into structured JSON.
4. ChatGPT presents the match and asks whether to create the CRM record.
5. On confirmation, ChatGPT calls create_company with the enriched fields.

You never opened Perplexity, never copied a URL, and never mapped fields manually. The MCP server handled the Perplexity integration, auth, and CRM write in one protocol layer. That is why first-party MCP support matters more than bolting a search bar onto a CRM UI.

Use Perplexity-powered company search when CRM records lack recent news, funding events, or leadership changes.

Perplexity's API documentation describes real-time search capabilities that complement static CRM enrichment.

Perplexity integration via search_company_ai complements static CRM records with current web intelligence (useful before client meetings when company news, funding rounds, or leadership changes affect your pitch. Run enrichment on CRM contacts first, then Perplexity search for company-level context the contact record does not capture.

Use Perplexity for company research before high-value meetings; use CRM enrichment for contact-level data) complementary, not redundant.

The Bottom Line