Lead Generation
Where do prospects come from, and how do you know they're worth pursuing?
Lead generation is not volume. It is the disciplined practice of finding people whose problems match what you've built, and earning the right to a conversation.
Overview
| Attribute | Value |
|---|---|
| Purpose | Surface qualified prospects who match ICP criteria |
| Trigger | Ongoing — pipeline below target, new market entry, or campaign launch |
| Frequency | Daily execution, weekly review, monthly strategy |
| Duration | Continuous — 2-4 hours/day for active prospecting |
| Owner | Sales Development (AI-led, human-reviewed) |
| Output | Scored prospect list in CRM with ICP fit evidence |
Human Role: ICP definition, channel strategy, relationship judgment AI Role: Prospect research, data enrichment, scoring, outreach drafting Spectrum: AI-Assisted (moving toward AI-Led)
Prerequisites
Tools Required
| Tool | Purpose | Access |
|---|---|---|
| CRM | Prospect database, pipeline tracking | Sales CRM |
| Prospect research, connection requests | Sales Navigator or free | |
| Email platform | Outreach delivery and tracking | Resend / similar |
| AI research | Prospect profiling, company analysis | Agent or manual |
| ICP document | Target definition with psycho-logic | ICP Framework |
Knowledge Requirements
- Validated ICP with psycho-logic profile (not just demographics)
- Understanding of target industry pain points
- Product value proposition mapped to ICP needs
- Competitive landscape awareness
Inputs
| Input | Source | Required? |
|---|---|---|
| ICP definition | ICP Workflow | Yes |
| Pipeline target | Sales strategy / quarterly plan | Yes |
| Current pipeline state | CRM dashboard | Yes |
| Past outreach data | CRM activity history | If available |
| Content assets | Marketing / knowledge base | Recommended |
Upstream Dependencies
| Upstream Workflow | What It Provides | Link |
|---|---|---|
| ICP Definition | Target archetype with psycho-logic profile | ICP |
| Market Sizing | TAM/SAM/SOM validation | Market Sizing |
| Content Pipeline | Berley assets for value-first outreach | Content Pipeline |
Process
Phase 1: Define Target
Duration: 1-2 hours (refresh quarterly) Responsibility: Human-led
Before generating any leads, confirm your target. Every prospecting session without a clear ICP wastes time on people who will never buy.
Step 1.1: Confirm ICP Criteria
- ICP document exists and is current (less than 90 days old)
- Psycho-logic profile complete (what they say vs what they do)
- Company-level filters defined (size, industry, geography, tech stack)
- Decision-maker roles identified (who signs, who influences, who blocks)
- Disqualification criteria defined (who is NOT a fit)
Step 1.2: Set Pipeline Targets
| Metric | Target | Current |
|---|---|---|
| Prospects researched per week | 20-30 | ? |
| Outreach sent per week | 15-20 | ? |
| Replies received per week | 3-5 (15%+ rate) | ? |
| Discovery calls booked per month | 5+ | ? |
Step 1.3: Choose Channels
Not all channels are equal. Start with one, prove it, then expand.
| Channel | Best For | Cost | Time to First Signal |
|---|---|---|---|
| B2B, decision-maker access | Free-$100/mo | 1-2 weeks | |
| Cold email | Scale, measurable | $20-50/mo | 1-2 weeks |
| Referrals | Highest conversion, lowest volume | Free | Ongoing |
| Content/inbound | Warm leads, longer cycle | Time investment | 1-3 months |
| Events/conferences | Face-to-face, high intent | $500-5000 | Immediate |
| Partnerships | Trusted introduction | Relationship | Variable |
Phase 1 Output: Confirmed ICP, pipeline targets, chosen channels
Phase 2: Research Prospects
Duration: 1-2 hours per batch of 10 Responsibility: AI-led, human-reviewed
This is where AI creates the most leverage. Research that took a salesperson 3 hours per prospect now takes minutes.
Step 2.1: Source Prospect List
| Source | Method | Volume |
|---|---|---|
| LinkedIn Sales Navigator | Search by ICP criteria | 50-100/search |
| Industry directories | Filter by sector, size, location | Variable |
| Conference attendees | Past event lists | 20-100/event |
| CRM existing contacts | Re-engage dormant leads | Variable |
| Content engagement | Who's reading your articles | Low but warm |
| Competitor reviews | Who's unhappy with alternatives | 5-10/week |
Step 2.2: Build Prospect Profile (AI-Led)
For each prospect, the agent researches and produces:
PROSPECT PROFILE
────────────────
Company: [Name]
Industry: [Sub-sector]
Size: [Employees / Revenue]
Location: [Geography]
Decision Maker: [Name, Title]
Pain Indicators:
- [Observable evidence of the problem you solve]
- [Technology they're using that you replace]
- [Recent events suggesting need (hiring, funding, expansion)]
ICP Fit Score: [0-100 based on criteria match]
Fit Evidence: [Specific reasons for the score]
Recommended Approach: [Channel + message angle]
Step 2.3: Human Review
Before any outreach, the human reviews the prospect list:
- Remove false positives (company looks right but isn't)
- Flag high-priority prospects for personal touch
- Identify warm introductions through network
- Verify no existing relationship conflicts
Phase 2 Output: Scored prospect list, 10-20 per batch, in CRM
Phase 3: ICP Validation
Duration: Ongoing — refine with every interaction Responsibility: Human judgment, AI data
This is the hard problem. A prospect can match every demographic filter and still not be a buyer. ICP validation is the bridge between "looks like a fit on paper" and "will actually buy."
Step 3.1: Match Against Psycho-Logic Profile
Demographics tell you WHO they are. Psycho-logic tells you WHETHER they'll buy.
| Signal Type | What to Look For | How to Find It |
|---|---|---|
| Stated pain | They describe the problem you solve | LinkedIn posts, interviews, reviews |
| Behavioral pain | They're working around the problem | Job postings (hiring for manual work), tech stack (spreadsheets, email) |
| Timing signal | Something changed recently | Funding round, leadership change, lost deal, regulation |
| Social proof | Peers in their industry already bought | Case studies, testimonials, public references |
| Budget authority | They can approve the spend | Title, org structure, company stage |
Step 3.2: Score ICP Fit
| Dimension | Weight | Score (0-5) | Evidence |
|---|---|---|---|
| Problem match | 30% | ? | Do they have the specific pain we solve? |
| Budget fit | 20% | ? | Can they afford it? Is there budget? |
| Authority | 20% | ? | Are we talking to the decision maker? |
| Need urgency | 15% | ? | Is this urgent or "nice to have"? |
| Timing | 15% | ? | Is now the right time (trigger event)? |
ICP Fit = weighted average. Above 70% = pursue. 40-70% = nurture. Below 40% = disqualify.
Step 3.3: Validate Through Interaction
The only true validation is conversation. No amount of research replaces hearing them describe their problem.
| Validation Stage | What You Learn | Method |
|---|---|---|
| Pre-contact | Does research match ICP? | AI research + human review |
| First touch | Do they respond? | Outreach response rate |
| First call | Do they name the pain unprompted? | Discovery call |
| Follow-up | Do they invest their time? | Meeting scheduling, info sharing |
| Proposal | Do they engage on specifics? | Questions, objections, negotiation |
The ICP validation loop:
RESEARCH → SCORE → OUTREACH → RESPONSE → CONVERSATION → VALIDATE
↑ |
└──── Update ICP profile with what you learned ───────────┘
Every conversation either confirms or challenges your ICP assumptions. Feed learnings back. The ICP is a living document.
Phase 3 Output: Validated prospects ready for outreach, updated ICP insights
Phase 4: Outreach
Duration: 30-60 minutes per batch Responsibility: AI drafts, human reviews and sends
Step 4.1: Compose Message (AI-Led)
Every outreach message must pass three tests:
- Relevant — References their specific situation, not generic pain
- Valuable — Offers insight or resource, not just a meeting request
- Human — Reads like a person wrote it, not a template
The Trojan Horse strategy: lead with value, not a pitch.
| Approach | When to Use | Example |
|---|---|---|
| Insight-led | They posted about a relevant problem | "Saw your post about RFP response time. We've been tracking how answer libraries compound..." |
| Value-led | You have content they'd benefit from | "Wrote this analysis of construction bid win rates — thought it might be useful given [context]" |
| Referral-led | Someone connected you | "[Name] mentioned you're dealing with [problem]. Happy to share what worked for [similar company]" |
| Event-led | Shared experience | "Good to hear your talk at [event]. Your point about [topic] aligns with something we've been building..." |
Step 4.2: Human Review Gate
Before any message is sent:
- Does this sound like me, not a robot?
- Is the personalization real (not "I noticed your company...")?
- Is there a clear value offering, not just "let's chat"?
- Would I reply to this if I received it?
Step 4.3: Multi-Touch Sequence
| Day | Channel | Action | Fallback |
|---|---|---|---|
| 0 | Connection request with note | If no LinkedIn, email | |
| 3 | Personalized outreach with value | — | |
| 7 | Engage with their content (like/comment) | — | |
| 10 | Follow-up with different angle | — | |
| 14 | Share relevant content, tag if appropriate | — | |
| 21 | Final touch — direct question | — | |
| 30 | — | Move to nurture list or disqualify | — |
Phase 4 Output: Outreach sent, tracked in CRM with activity logging
Phase 5: Measure and Learn
Duration: 30 minutes weekly Responsibility: AI reports, human interprets
Step 5.1: Track Metrics
| Metric | Target | Actual | Diagnosis if Below |
|---|---|---|---|
| Prospects researched | 20/week | ? | Increase research volume or improve sources |
| Outreach sent | 15/week | ? | Remove bottleneck in review gate |
| Response rate | >15% | ? | Message quality or ICP mismatch |
| Discovery calls | 5/month | ? | Qualification criteria too loose/tight |
| Pipeline value added | $X/month | ? | Deal sizing or volume issue |
Step 5.2: Channel Effectiveness
| Channel | Sent | Replied | Meeting | Cost per Meeting |
|---|---|---|---|---|
| ? | ? | ? | ? | |
| ? | ? | ? | ? | |
| Referral | ? | ? | ? | ? |
| Inbound | N/A | N/A | ? | ? |
The Explore-Exploit algorithm optimizes channel allocation: explore new channels with 20% of effort, exploit proven channels with 80%.
Step 5.3: ICP Refinement
After every batch of outreach:
- Which prospects responded? What do they share?
- Which prospects didn't? What disqualifies them?
- Update ICP fit criteria based on response patterns
- Identify new segments that responded unexpectedly
Phase 5 Output: Updated metrics, channel performance, ICP refinements
Outputs
| Output | Format | Destination |
|---|---|---|
| Scored prospect list | CRM contacts with ICP fit score | CRM Contacts |
| Outreach activity log | CRM activities per contact | CRM Activity Timeline |
| Channel performance | Weekly metrics report | Sales review |
| ICP refinements | Updated ICP document | ICP Framework |
Downstream Consumers
| Downstream Workflow | What It Needs | Link |
|---|---|---|
| Lead Qualification | Scored prospects with research profiles | Qualification |
| Funnel Engineering | New leads entering pipeline | Funnel |
| Sales Dev Agent | Prospect data for agent outreach | PRD |
Success Criteria
Quality Metrics
| Metric | Target | Measurement |
|---|---|---|
| ICP fit accuracy | >70% of qualified leads match ICP | Post-call validation |
| Response rate | >15% on outreach | CRM tracking |
| Research-to-outreach ratio | >80% of researched prospects receive outreach | Pipeline flow |
| False positive rate | <30% (prospects who looked good but weren't) | Disqualification tracking |
Performance Metrics
| Metric | Target | Timeframe |
|---|---|---|
| Discovery calls booked | 5+/month | Monthly |
| Pipeline value generated | >$50K/month from new leads | Monthly |
| Time from research to first touch | <48 hours | Per prospect |
| Cost per qualified lead | <$50 | Monthly average |
Failure Modes
| Failure | Symptom | Diagnosis | Solution |
|---|---|---|---|
| Volume without quality | High outreach, zero replies | ICP mismatch or generic messages | Review ICP, personalize outreach |
| Research paralysis | Perfect profiles, no outreach sent | Fear of rejection or perfectionism | Set "good enough" threshold, send |
| Channel fixation | Only using one channel | 100% LinkedIn, 0% email | Explore-exploit: 20% to new channels |
| Stale pipeline | Same prospects for weeks, no new intake | Research dried up or targets exhausted | Expand ICP, new sources, new segment |
| AI dependency | Agent sends, human never reviews | Generic messages, reputation damage | Enforce human review gate on every send |
| No feedback loop | Outreach goes out, nothing comes back | No tracking, no learning | Wire CRM activity logging, weekly review |
The Trojan Horse
The highest-leverage lead gen strategy: give away something so valuable that prospects come to you.
| Asset Type | Example | Why It Works |
|---|---|---|
| Industry benchmark | "Construction bid win rates by company size" | They see where they stand |
| Diagnostic tool | "Score your RFP process maturity" | Self-identified pain |
| Case study | "How [company] cut RFP time by 70%" | Social proof + specific numbers |
| Template | "RFP response template with auto-scoring" | Immediate utility |
| Community | Industry-specific peer group | Belonging + ongoing access |
You can have everything in life you want, if you will just help enough other people get what they want — Zig Ziglar
Psychological Biases that power the Trojan Horse:
- Reciprocity — Give value first, ask second
- Consistency — Once they engage, they're more likely to continue
- Social proof — Show others like them already using it
Context
- Lead Qualification — What happens after a lead is generated
- ICP Framework — How to define WHO to target
- Sales Dev Agent — AI agent that executes this workflow
- Sales Work Chart — Human/AI split for all sales jobs
- Process Optimisation — How to improve this workflow over time
- Behavioural Biases — Psychology that drives buyer behavior