Search for leads by talking: the Lode Leads conversational agent

Every lead generation tool makes you think in their language. You pick filters from dropdowns. You select industries from a taxonomy someone else built. You write boolean strings and hope you did not miss a synonym.

Lode Leads works differently. You describe what you want in plain English — inside the dashboard, or directly from ChatGPT or Claude via our native MCP integration — and the AI agent does the rest.

How it works

The conversational agent lives inside the Lode Leads dashboard and is also exposed as a native MCP server, which means you can run the same campaigns directly from ChatGPT, Claude, or any MCP-compatible AI assistant. Instead of clicking through filter menus, you type what you are looking for:

"Find marketing directors at e-commerce companies in the UAE with 50-200 employees. They should be active on LinkedIn and ideally also have an Instagram business presence."

The agent parses your request, identifies the key parameters, and configures a per-platform campaign automatically:

• Intent classification: The agent determines whether you are looking for B2B leads, B2C profiles (creators or influencers), or local businesses (Google Maps). That drives which platform you should target first.
• Parameter extraction: From your natural language description, the agent extracts job titles, industries, company sizes, locations, platform preferences, and qualifying criteria.
• Platform selection and campaign creation: Each LodeLeads campaign targets one platform — the agent recommends the right one for your ICP and sets up the filters. If you want LinkedIn and Instagram coverage, it creates two campaigns on the same account and deduplicates the combined lead list.
• Results delivery: Leads are returned with AI scores, platform profile data and outreach angles where available, and company details. The agent can explain why specific leads scored high or low.

Natural language beats filter menus

Filter-based tools force you to decompose your mental model of a good lead into discrete field values. This creates three problems:

• You miss nuance. "Companies that are hiring aggressively" is not a dropdown option. But it is a signal the AI can detect from job posting data and headcount changes.
• You miss combinations. "Restaurants in Cairo with good Google reviews but no Instagram presence" requires cross-referencing two platforms. In a conversation, it is one sentence.
• You miss what you do not know to ask for. The agent can suggest additional qualifying criteria: "You mentioned e-commerce companies. Should I also filter for companies that sell internationally, since your product is a shipping solution?"

Per-platform campaign orchestration

A LodeLeads campaign targets one platform at a time — LinkedIn, Google Maps, TikTok, Instagram, Facebook, Twitter/X, Reddit, or YouTube. Based on your request, the agent recommends the platform most likely to surface your ICP and launches the campaign. If your ICP lives across multiple platforms, you run multiple campaigns on the same account, the same credit pool, and LodeLeads deduplicates leads that appear on more than one.

Refinement through conversation

After initial results come back, you refine through conversation:

• "The scores are good but I'm only seeing companies in Dubai. Expand to Riyadh and Jeddah."
• "These are too large. Filter to companies under 100 employees."
• "Show me more like the third result -- that is exactly the profile I want."
• "Export the top 20 leads with a score above 75 to Excel."

The agent maintains context throughout the conversation. You do not need to repeat your base criteria for every adjustment.

Get early access to the conversational agent — available inside the dashboard and as a native MCP server for ChatGPT and Claude. Describe your ICP in one sentence and get filtered, verified leads worth dialing. Get started free.