Skip to main content

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

AttributeValue
PurposeSurface qualified prospects who match ICP criteria
TriggerOngoing — pipeline below target, new market entry, or campaign launch
FrequencyDaily execution, weekly review, monthly strategy
DurationContinuous — 2-4 hours/day for active prospecting
OwnerSales Development (AI-led, human-reviewed)
OutputScored 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

ToolPurposeAccess
CRMProspect database, pipeline trackingSales CRM
LinkedInProspect research, connection requestsSales Navigator or free
Email platformOutreach delivery and trackingResend / similar
AI researchProspect profiling, company analysisAgent or manual
ICP documentTarget definition with psycho-logicICP 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

InputSourceRequired?
ICP definitionICP WorkflowYes
Pipeline targetSales strategy / quarterly planYes
Current pipeline stateCRM dashboardYes
Past outreach dataCRM activity historyIf available
Content assetsMarketing / knowledge baseRecommended

Upstream Dependencies

Upstream WorkflowWhat It ProvidesLink
ICP DefinitionTarget archetype with psycho-logic profileICP
Market SizingTAM/SAM/SOM validationMarket Sizing
Content PipelineBerley assets for value-first outreachContent 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

MetricTargetCurrent
Prospects researched per week20-30?
Outreach sent per week15-20?
Replies received per week3-5 (15%+ rate)?
Discovery calls booked per month5+?

Step 1.3: Choose Channels

Not all channels are equal. Start with one, prove it, then expand.

ChannelBest ForCostTime to First Signal
LinkedInB2B, decision-maker accessFree-$100/mo1-2 weeks
Cold emailScale, measurable$20-50/mo1-2 weeks
ReferralsHighest conversion, lowest volumeFreeOngoing
Content/inboundWarm leads, longer cycleTime investment1-3 months
Events/conferencesFace-to-face, high intent$500-5000Immediate
PartnershipsTrusted introductionRelationshipVariable

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

SourceMethodVolume
LinkedIn Sales NavigatorSearch by ICP criteria50-100/search
Industry directoriesFilter by sector, size, locationVariable
Conference attendeesPast event lists20-100/event
CRM existing contactsRe-engage dormant leadsVariable
Content engagementWho's reading your articlesLow but warm
Competitor reviewsWho's unhappy with alternatives5-10/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 TypeWhat to Look ForHow to Find It
Stated painThey describe the problem you solveLinkedIn posts, interviews, reviews
Behavioral painThey're working around the problemJob postings (hiring for manual work), tech stack (spreadsheets, email)
Timing signalSomething changed recentlyFunding round, leadership change, lost deal, regulation
Social proofPeers in their industry already boughtCase studies, testimonials, public references
Budget authorityThey can approve the spendTitle, org structure, company stage

Step 3.2: Score ICP Fit

DimensionWeightScore (0-5)Evidence
Problem match30%?Do they have the specific pain we solve?
Budget fit20%?Can they afford it? Is there budget?
Authority20%?Are we talking to the decision maker?
Need urgency15%?Is this urgent or "nice to have"?
Timing15%?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 StageWhat You LearnMethod
Pre-contactDoes research match ICP?AI research + human review
First touchDo they respond?Outreach response rate
First callDo they name the pain unprompted?Discovery call
Follow-upDo they invest their time?Meeting scheduling, info sharing
ProposalDo 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:

  1. Relevant — References their specific situation, not generic pain
  2. Valuable — Offers insight or resource, not just a meeting request
  3. Human — Reads like a person wrote it, not a template

The Trojan Horse strategy: lead with value, not a pitch.

ApproachWhen to UseExample
Insight-ledThey posted about a relevant problem"Saw your post about RFP response time. We've been tracking how answer libraries compound..."
Value-ledYou have content they'd benefit from"Wrote this analysis of construction bid win rates — thought it might be useful given [context]"
Referral-ledSomeone connected you"[Name] mentioned you're dealing with [problem]. Happy to share what worked for [similar company]"
Event-ledShared 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

DayChannelActionFallback
0LinkedInConnection request with noteIf no LinkedIn, email
3EmailPersonalized outreach with value
7LinkedInEngage with their content (like/comment)
10EmailFollow-up with different angle
14LinkedInShare relevant content, tag if appropriate
21EmailFinal touch — direct question
30Move 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

MetricTargetActualDiagnosis if Below
Prospects researched20/week?Increase research volume or improve sources
Outreach sent15/week?Remove bottleneck in review gate
Response rate>15%?Message quality or ICP mismatch
Discovery calls5/month?Qualification criteria too loose/tight
Pipeline value added$X/month?Deal sizing or volume issue

Step 5.2: Channel Effectiveness

ChannelSentRepliedMeetingCost per Meeting
LinkedIn????
Email????
Referral????
InboundN/AN/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

OutputFormatDestination
Scored prospect listCRM contacts with ICP fit scoreCRM Contacts
Outreach activity logCRM activities per contactCRM Activity Timeline
Channel performanceWeekly metrics reportSales review
ICP refinementsUpdated ICP documentICP Framework

Downstream Consumers

Downstream WorkflowWhat It NeedsLink
Lead QualificationScored prospects with research profilesQualification
Funnel EngineeringNew leads entering pipelineFunnel
Sales Dev AgentProspect data for agent outreachPRD

Success Criteria

Quality Metrics

MetricTargetMeasurement
ICP fit accuracy>70% of qualified leads match ICPPost-call validation
Response rate>15% on outreachCRM tracking
Research-to-outreach ratio>80% of researched prospects receive outreachPipeline flow
False positive rate<30% (prospects who looked good but weren't)Disqualification tracking

Performance Metrics

MetricTargetTimeframe
Discovery calls booked5+/monthMonthly
Pipeline value generated>$50K/month from new leadsMonthly
Time from research to first touch<48 hoursPer prospect
Cost per qualified lead<$50Monthly average

Failure Modes

FailureSymptomDiagnosisSolution
Volume without qualityHigh outreach, zero repliesICP mismatch or generic messagesReview ICP, personalize outreach
Research paralysisPerfect profiles, no outreach sentFear of rejection or perfectionismSet "good enough" threshold, send
Channel fixationOnly using one channel100% LinkedIn, 0% emailExplore-exploit: 20% to new channels
Stale pipelineSame prospects for weeks, no new intakeResearch dried up or targets exhaustedExpand ICP, new sources, new segment
AI dependencyAgent sends, human never reviewsGeneric messages, reputation damageEnforce human review gate on every send
No feedback loopOutreach goes out, nothing comes backNo tracking, no learningWire 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 TypeExampleWhy 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
CommunityIndustry-specific peer groupBelonging + 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