← Back to Blog

AI Agent for Lead Generation: From Web Research to Telegram in One Prompt

Most sales teams spend 40% of their week on prospecting — finding the right people, qualifying them, building lists. An AI agent can do that work in minutes, while you focus on closing.

Published by GetClawCloud · May 2, 2026

Here's the truth about B2B lead generation: the bottleneck isn't your pitch. It's finding who to pitch to. The person who needs your product is out there — on LinkedIn, on Crunchbase, on industry forums, at the right company, with the right title, at the right growth stage. Finding them is manual, repetitive, and soul-crushing after the 50th tab-open.

Tools like Apollo and ZoomInfo exist, but they cost hundreds a month and still require human judgment to filter, qualify, and prioritise. CRMs like Salesforce can automate sequences but can't do the discovery — the part where you figure out who's worth targeting in the first place.

An AI lead generation agent bridges this gap. It searches the web, evaluates fit, enriches profiles, and delivers qualified leads directly to your Telegram. Configure it once. Let it run on a schedule. Wake up to a list of companies and people who match your ideal customer profile.

        The best sales teams don't get more leads — they get better signals. An AI agent that filters noise before you see it is worth more than a thousand cold email addresses.
      

What an AI Lead Generation Agent Actually Does

Not all "lead gen automation" is the same. Here's what a well-designed agent pipeline looks like:

Stage 1: Discovery
The agent searches across multiple sources — Crunchbase for recent funding announcements, LinkedIn for companies hiring in specific roles, product review sites for users complaining about solving the problem you sell to, industry news for expansion announcements, job boards for relevant roles. It covers ground a human could never scan in one sitting.

Stage 2: Qualification
Each candidate is scored against your Ideal Customer Profile (ICP). Company size, industry, geography, technology stack, recent signals (funding, hiring, expansion), and decision-maker titles. The agent rejects firms with weak alignment and ranks the rest by fit strength.

Stage 3: Enrichment
For each qualified lead, the agent enriches the profile with: key decision-maker names and LinkedIn URLs, recent company news, technology stack (from job postings and product pages), and a brief "angle" — why this company might need your product right now.

Stage 4: Delivery
The agent compiles everything into a clean Telegram briefing — sorted by priority, actionable, and ready for outreach. No spreadsheet, no dashboard login.

The Prompt: Your AI Lead Generation Agent

The prompt below turns any OpenClaw-powered Telegram bot into a dedicated lead generation engine. Copy it, send it to your bot, then tell it what you sell and who you want to target.

How to use:

Deploy OpenClaw on GetClawCloud — one click, no server setup
Connect Telegram (built-in, one-click pairing)
Paste this prompt as your first message
Follow up with your product pitch, ideal customer profile, and any targeting criteria

You are an AI Lead Generation Agent. Your job is to find, qualify, and enrich sales leads — all from public web sources. ## Workflow ### Phase 1: Setup Ask the user for: 1. What they sell (product/service, category, price range) 2. Their Ideal Customer Profile (ICP): - Target industries - Company size range (employees or revenue) - Geography - Decision-maker titles - Technology stack or integration requirements (if any) 3. Buying signals to prioritise (e.g., recent funding, executive hires, expansion, API purchase, competitor usage) 4. Any companies to exclude (existing customers, competitors, irrelevant segments) 5. Preferred output format (compact briefing vs detailed profiles) 6. Schedule preference (one-time list or recurring scans) Confirm the setup before proceeding. ### Phase 2: Discovery (for each lead generation session) Use web search to find matching leads across multiple sources: 1. **Recent funding** — "recent Series [A/B/C] funding [industry] [region] last 30 days" 2. **Growth signals** — "hiring [role relevant to your product] [industry]" 3. **Pain points** — "[problem your product solves] review OR complaint OR alternative" 4. **Company expansions** — "[industry] new office OR expansion OR partnership [region]" 5. **Technology research** — "companies evaluating [relevant tech] OR [competitor product] alternative" 6. **Job postings** — search for roles that indicate the company needs your solution For each source, fetch and evaluate the actual page — don't rely solely on search snippets. ### Phase 3: Qualification & Scoring For each candidate company, analyse: - Company size and industry (does it match ICP?) - Location (in target geography?) - Recent events (funding, hiring, product launches) - Technology adoption (do they use tools that yours complements or replaces?) - Decision-maker accessibility (are the right titles present on LinkedIn or visible in job posts?) Score each lead: HOT (high fit + buying signal), WARM (good fit, no immediate signal), COLD (partial fit, lower priority). ### Phase 4: Enrichment For HOT and WARM leads, enrich with: - Company name and website - Key decision-makers with LinkedIn URLs (when findable from public sources) - Recent news relevant to why they might buy now (1 sentence) - The "angle" — a specific reason they'd need your product today - Company tech stack clues (from job postings, product pages, integrations pages) ### Phase 5: Delivery Present results in this format: **📋 Lead Briefing — [Date]** **🔥 Hot Leads (Act Now)** 1. Company A (Size, Industry, Location) - Signal: [what they did that makes them a hot lead] - Angle: [how your product fits their situation] - Decision-makers: [Name, Title, LinkedIn URL] - 🔗 [Company URL] **✅ Warm Leads (Nurture)** [List with same structure but shorter] **📎 Full Sources** [URLs for each significant source] ## Rules - Only use public, verifiable sources — no scraping gated databases - Never invent contact details or email addresses - Flag uncertainty (e.g., "LinkedIn profile found, title inferred from job post") - Prioritise quality over quantity: 3 hot leads beat 20 weak ones - Always include the reasoning behind each score - Output in Telegram-friendly format (bold for emphasis, no tables, bullet lists) - For recurring scans: flag leads that are NEW and leads that have CHANGED status ## Start Ask the user for their product and ICP to begin.

💡 Works with any OpenClaw agent with web search access. No API keys beyond your OpenClaw setup needed.

Real Output: What Shows Up in Telegram

Configured for a hypothetical SaaS that sells AI-powered contract analysis to legal teams, the agent might deliver:

📋 Lead Briefing — May 2, 2026

🔥 Hot Leads

1. LexAI Corp (Series B, 150-300 employees, San Francisco)
Signal: Just raised $25M Series B for AI legal tools. Hiring 3 contract management roles.
Angle: Expanding their AI legal stack. Your contract analysis integrates with their existing workflow.
Decision-maker: Sarah Chen, VP Legal Ops — linkedin.com/in/sarahlchen
🔗 lexaicorp.com

2. Legacy Legal LLP (500+ employees, London)
Signal: Hired a Head of Legal Innovation last month. Published a blog post about digitising contract workflows.
Angle: Actively modernising. Your product fills a gap they've publicly acknowledged.
🔗 legacylegal.co.uk

✅ Warm Leads

3. DocuSafe Inc (Mid-market, Chicago)
Signal: Launched a legal document portal. No direct AI contract analysis yet — potential upsell.
Decision-maker: Not publicly visible. Try general counsel@docusafe.com or LinkedIn search.

Every lead includes the reasoning, not just the name. You can decide whether to pursue, adjust your ICP, or refine the agent's targeting — all from one Telegram conversation.

Scheduling: Make It Run Every Day

One lead scan is useful. A daily drip of fresh, qualified leads is transformative.

Set up recurring lead generation with OpenClaw cron:


# Run daily lead scan at 6 AM UTC
openclaw cron add --every 24h --text "Run lead generation agent. Search for new leads matching the ICP configured earlier. Flag new companies and any changes in existing leads."

The briefing delivers to your Telegram automatically. Your phone buzzes with fresh leads before you pour your morning coffee.

Who This Actually Helps

Solo founders & early-stage startups — generate leads with zero SDR headcount
Small B2B sales teams (1-5 people) — double pipeline output without doubling headcount
Freelancers & agencies — find project opportunities without spending hours on Upwork
GTM teams — monitor who's entering your market and prepare land-grab outreach
Venture investors — identify portfolio-adjacent companies that need your solution

Why This vs. Existing Lead Gen Tools

Apollo/ZoomInfo: $99-$500+/mo. Great databases but you still need human judgment to qualify. No built-in signal detection — you pay for raw contacts and sort them yourself.

LinkedIn Sales Navigator: $99+/mo. The best social platform for prospecting — but you spend hours filtering, building lists, and exporting. Manual by design.

AI Lead Gen Agent (this approach): Runs on OpenClaw's free tier. Searches across sources (not just one database). Qualifies, enriches, and ranks. Delivers to Telegram. You configure once and get results on autopilot.

        The tools industry solution was "bigger database, more contacts." The AI solution is "smarter filtering, fewer but better contacts." The lead generation agent is the second approach — and it costs effectively nothing to run.
      

Refining Your Agent Over Time

The beauty of a prompt-based agent is iteration. After your first run, you'll notice things:

"Too many companies in the wrong geography" → narrow the ICP location field
"These leads are too early-stage" → exclude companies under 50 employees or pre-revenue
"I need vertical-specific signals" → add custom buying signals like "HIPAA compliance" or "SOC2 audit in progress"
"Too much noise per briefing" → raise the score threshold so only HOT leads trigger Telegram notifications

Each adjustment takes 30 seconds — edit the setup instructions and re-run. The agent adapts immediately.

Getting Started in 2 Minutes

Deploy OpenClaw on GetClawCloud — one click, zero server configuration
Connect Telegram — native integration, no webhooks or middleware
Paste the prompt above, then describe your product and ideal customer
Schedule it with cron for daily lead delivery

Your first qualified lead arrives the next time the agent runs. Adjust the prompt as your ICP evolves — the agent grows with you.

        Lead generation is a data problem dressed up as a sales problem. An AI agent that collects, qualifies, and ranks data solves the real bottleneck. Copy the prompt, paste it into your Telegram bot, and see what surfaces by tomorrow morning.
      

Launch Your AI Lead Generation Agent

Deploy OpenClaw with one click, connect Telegram, and paste the lead gen prompt. Qualified leads delivered to your phone on autopilot.

Start on GetClawCloud →