AI SDR Workflow

Build a complete AI sales development system via OpenClaw Ultra. From defining your ICP and researching target accounts to running multi-channel outreach sequences, classifying replies, qualifying leads, and handing off hot prospects to your closing team, manage your entire SDR operation from a single chat interface.

Core System Overview

INFO

This is a closed-loop AI SDR workflow. OpenClaw Ultra handles lead ingestion, account research, message personalization, sequence execution, reply triage, lead qualification, meeting booking, and CRM handoff, so you can generate qualified pipeline continuously without switching between prospecting tools.

System Layer	Core Function	Final Output
Ingestion Layer	ICP definition, lead source integration, list building	Scored and segmented target account list
Research Layer	Account enrichment, trigger event detection, pain hypothesis	Researched account profiles with hooks
Outreach Layer	Multi-channel message generation, cadence design	Channel-specific outreach sequence
Execution Layer	Cross-channel sending, engagement tracking, adaptive follow-up	Executed campaign with signal capture
Reply Layer	Inbound triage, intent classification, FAQ response	Categorized reply log
Qualification Layer	BANT scoring, fit + intent + timing scoring, disqualification rules	Scored leads with routing decision
Handoff Layer	Meeting booking, CRM sync, handoff documentation	Qualified leads passed to sales with full context
Optimization Layer	Performance analysis, ICP refinement, message iteration	Improved conversion rates next cycle

Prerequisites

Item	Requirement
OpenClaw Ultra	Installed and running
ICP Defined	Industry, company size, role, and geography for your ideal prospect
Lead Source	Database, CSV list, or CRM export of target accounts
CRM Account	(Recommended) Active account with API access for lead handoff
Sending Domains	(Recommended) 1-2 secondary domains with SPF/DKIM/DMARC configured for cold email
Calendar Access	(Recommended) Google Calendar or similar for automated meeting booking

Step 0 — Initialize Your AI SDR System

Set up OpenClaw Ultra as your dedicated AI sales development representative.

Operation Steps

1. Open OpenClaw Ultra new chat session
2. Prepare your ICP definition, target account list, and sending domains
3. Paste the initialization prompt

Ready-to-Use Prompt

Act as my AI sales development representative.

My ICP:
- Industry: [industry]
- Company size: [size range]
- Target roles: [job titles]
- Geography: [regions]
- Ideal customer profile: [brief description]
- Lead source: [database, CSV, or CRM]
- Sending domains: [domain1.com, domain2.com]

Build a complete AI SDR system that covers:
- lead ingestion and account scoring
- account research and trigger event detection
- multi-channel outreach creation (email + LinkedIn)
- sequence execution and engagement tracking
- reply triage and lead qualification
- meeting booking and CRM handoff

Step 1 — Define Your ICP and Ingest Leads

Set the targeting criteria and bring your account list into the system.

1.1 Define Your Ideal Customer Profile

Define my target account criteria for this SDR campaign:

Target parameters:
- Industry: [industry or industries]
- Company size: [employee count or revenue range]
- Departments: [departments to target]
- Job titles: [specific roles]
- Geography: [countries or regions]
- Tech stack: [tools they use, if relevant]
- Buying triggers: [events that signal intent — funding, hiring, product launch]

Return a structured ICP document with:
- Primary targets (best fit)
- Secondary targets (good fit)
- Exclusion criteria (who not to pursue)
- Buying committee map for each target account type
- Trigger event priority list ranked by conversion likelihood

1.2 Ingest and Score Your Lead List

Import and score my target account list against the ICP:

Lead source: [CSV / CRM export / database name]
Columns available: [list of columns in your data]

For each account:
- Score ICP fit from 1-5 based on firmographic match
- Score data completeness from 1-5 (missing contact info, empty fields)
- Flag accounts with buying triggers in the last 90 days
- Assign priority tier: hot (score 9-10), warm (score 6-8), explore (score below 6)

Return a prioritized account table sorted by combined score:
| Account | ICP Fit | Data Quality | Triggers | Priority | Next Action |

Step 1 Output

Prioritized target account list with ICP scores and buying trigger flags.

Step 2 — Research Accounts and Find Angles

Deep research each target account to build personalized outreach angles.

2.1 Research Target Accounts

Research the top-priority accounts from my list:

Accounts to research: [list of account names or domains]

For each account, gather:
- Company description, recent news, and funding history (last 6 months)
- Leadership changes, hiring trends, and org structure
- Tech stack they use and recent technology changes
- Public announcements (earnings, product launches, partnerships)
- Key decision-makers with LinkedIn profiles and role context

Return a research brief for each account with:
- Company snapshot (industry, size, revenue trend)
- Recent trigger events with dates
- Decision-maker table (name, title, tenure, LinkedIn URL)
- Personalization hooks ranked by relevance

2.2 Generate Pain Hypotheses and Outreach Angles

For each researched account, generate outreach angles:

Account research briefs: [paste from 2.1]

For each account:
- Identify 2-3 potential pain points based on their industry, size, and recent events
- Map each pain point to your value proposition
- Generate a personalized hook for the first-touch message
- Rate each angle by likely relevance (high / medium / low)

Return an angle brief per account:
| Account | Pain Hypothesis | Relevant Trigger | Hook Suggestion | Angle Score |

Step 2 Output

Research briefs with trigger events, personalization hooks, and ranked outreach angles per account.

Step 3 — Build Multi-Channel Outreach Sequences

Create personalized sequences that coordinate email, LinkedIn, and optional phone touches.

3.1 Write Channel-Specific First-Touch Messages

Write first-touch messages for my outreach campaign:

Target account: [account name]
Decision maker: [name, title]
Personalization hook: [trigger event or pain hypothesis from step 2]
Value proposition: [what we offer]

Create:
1. Cold email (under 120 words, subject line under 50 characters, one CTA)
2. LinkedIn connection request note (under 200 characters, conversational)
3. LinkedIn message (for after connection accepted, under 80 words)

Requirements:
- Reference the personalization hook in the first sentence of each channel
- No attachments or links in the first message
- CTA should be low-friction ("quick question" or "15-minute call", not a demo request)

Return each message with [bracketed placeholders] for custom fields.

3.2 Build an Adaptive Multi-Channel Cadence

Build a 21-day multi-channel outreach sequence for my target accounts:

Channels: email + LinkedIn (phone optional)

Cadence structure:
- Day 1: Email — personalized intro with hook
- Day 3: LinkedIn — send connection request with personalized note
- Day 7: Email — follow-up with new angle or industry insight
- Day 10: LinkedIn message — after connection accepted, short value-add
- Day 14: Email — case study or social proof angle
- Day 21: Email — breakup message, leave the door open

Adaptive rules:
- If prospect replies at any point, exit sequence and route to reply handling
- If prospect opens email 3+ times without replying, trigger a priority follow-up
- If LinkedIn connection is not accepted within 7 days, skip LinkedIn touches
- If email bounces, remove from sequence and flag for data verification

Return the full sequence with timing, channel, message content, and branching rules.

Step 3 Output

Multi-channel outreach sequence with adaptive branching rules, ready for deployment.

Step 4 — Execute Sequences and Track Engagement

Launch your outreach campaigns and monitor prospect engagement signals.

4.1 Launch Account Sequences

Launch outreach sequences for my prioritized accounts:

Campaign name: [name]
Accounts to sequence: [select accounts from step 1 priority list]
Sequence: [select sequence from step 3]
Daily send limit: [number per day across all domains]
Sending domains: [domain1.com, domain2.com]

Tracking: opens, clicks, replies, bounces, unsubscribe requests
LinkedIn: connection acceptance, message replies

Return a launch confirmation with:
- Accounts queued by priority tier
- Channels activated
- Estimated completion date
- Key metrics dashboard columns

4.2 Monitor Engagement Signals

Check my campaign [campaign name] engagement status:

Report:
- Emails sent / delivered / opened / replied
- LinkedIn connections accepted / messages replied
- Bounce rate (flag if above 3%)
- Unsubscribe rate (flag if above 0.5%)
- Positive reply rate
- Accounts with 3+ opens but no reply (priority follow-up candidates)

Return an engagement dashboard with:
- Green: healthy metrics
- Yellow: watch items requiring attention
- Red: pause campaign segment and investigate

Flag any accounts that need immediate human attention based on engagement signals.

Step 4 Output

Running campaign with live engagement tracking and signal-based alerting.

Step 5 — Classify Replies and Qualify Leads

Process incoming replies, triage intent, and score leads for handoff readiness.

5.1 Triage Inbound Replies

Review replies from campaign [campaign name]:

For each reply, classify intent:
- Positive: asked for call, demo, or pricing
- Question: asked about features, implementation, or use case fit
- Neutral: acknowledged but non-committal ("not now", "send more info")
- Negative: declined, not interested, wrong person
- Out of office: auto-reply — pause sequence for [X] days then retry
- Unsubscribe: remove from all sequences immediately
- FAQ: common question that can be auto-answered from knowledge base

For positive and question replies:
- Summarize the request in one sentence
- Suggest the next action and timeline
- Flag if the reply contains budget, authority, or timeline signals
- Auto-answer FAQ replies from approved response templates

Return a categorized reply log sorted by priority, with suggested next action for each.

5.2 Score Leads for Handoff

Score these engaged leads for sales handoff:

Leads to score: [list of prospects who replied positively or asked qualifying questions]

Score each on three dimensions:
- ICP fit (1-5): firmographic match to target profile
- Intent strength (1-5): specificity of their reply, buying signals detected
- Timing (1-5): mentioned timeline, urgency signals, recent trigger events

Combine into a composite score (fit + intent + timing, max 15).

Routing rules:
- Score 12-15: route to sales immediately — qualified and urgent
- Score 8-11: enter 14-day nurture with value-add content, re-score after engagement
- Score below 8: move to monthly nurture, suppress from active sequences for 90 days
- Disqualified: flag with reason, log for ICP refinement

Return a scored lead table:
| Lead | Account | Fit | Intent | Timing | Total | Route | Next Step |

Step 5 Output

Scored and routed leads: qualified leads to sales handoff, warm leads to nurture, cold leads suppressed.

Step 6 — Book Meetings and Hand Off to Sales

Convert qualified leads into booked meetings and hand off with full context to the closing team.

6.1 Book Qualified Meetings

Book meetings for leads scored 12+ and flagged for handoff:

Leads to book: [list of qualified leads with scores and reply context]

For each lead:
- Suggest 2-3 meeting time slots based on their reply context
- Generate a calendar invite with call agenda based on their expressed interest
- Send confirmation with call details, agenda, and prep notes
- Log the booking in CRM with campaign source attribution

Return a booking confirmation table:
| Lead | Proposed Times | Agenda | Calendar Sent | CRM Logged |

6.2 Generate Sales Handoff Summaries

Generate a handoff document for each booked meeting:

Lead: [name, company, title]

Handoff summary:
- Outreach history: channels used, messages sent, replies received, timing
- Qualification summary: score breakdown (fit + intent + timing), key signals detected
- What they responded to: specific hook, message, or angle that worked
- Their expressed interest: what they asked about or said they need
- Recommended call agenda: opening topic based on their reply history
- Objections raised: any concerns mentioned and suggested responses
- Relationship map: other stakeholders at the account worth involving

Format as a clean summary the AE can read in 3 minutes before the call.

Step 6 Output

Booked meetings on the calendar with complete handoff summaries for the sales team.

Step 7 — Analyze Performance and Optimize

Measure campaign results, identify what worked, and refine your SDR strategy for the next cycle.

7.1 Generate Campaign Performance Report

Generate a performance report for SDR campaign [campaign name]:

Metrics:
- Total accounts sequenced
- Reply rate by channel (email vs LinkedIn)
- Positive reply rate (interested / qualified)
- Meeting booking rate from positive replies
- Show rate for booked meetings
- Pipeline value generated from qualified leads
- Cost per qualified lead
- Bounce rate and spam complaint rate
- Sequence completion rate
- Average time from first touch to meeting booked

Compare to previous campaign [previous campaign name] if available. Flag top 3 metrics that improved and top 3 that declined.

7.2 Analyze Channel and Message Performance

Analyze which channels, messages, and angles performed best:

Channel performance:
- Email: open rate, reply rate, bounce rate
- LinkedIn: connection acceptance rate, message reply rate
- Multi-channel vs single-channel: meeting book rate comparison

Message performance:
- Which subject lines had the highest open rates
- Which hooks and angles generated the most positive replies
- Which CTA had the highest meeting conversion rate

Angle performance:
- Which trigger events produced the highest reply rates (funding, hiring, product launch)
- Which pain hypotheses resonated most by industry segment

Return a ranking table of what to keep, what to change, and what to drop for the next campaign.

7.3 Refine ICP and Targeting Rules

Based on campaign results, update my targeting rules:

Won meeting analysis:
- Common industries, company sizes, and titles that converted
- Trigger events that preceded booking
- Objections that prevented booking
- Sales cycle length from first touch to meeting

Silent and disqualified analysis:
- Patterns in accounts that never engaged
- Common early disqualification signals
- ICP segments that produced no positive replies

Return updated targeting rules:
- Include: [criteria that produced qualified meetings]
- Exclude: [criteria that produced no engagement or disqualified leads]
- Test: [new segments or trigger signals worth exploring next campaign]
- Updated ICP document with real conversion data

Step 7 Output

Campaign performance data driving refined ICP, improved messaging, and better targeting for the next SDR cycle.

Final Closed-Loop AI SDR Workflow

ICP Defined → Accounts Prioritized → Research → Sequences Built →
Campaigns Executed → Replies Classified → Leads Scored →
Meetings Booked → Performance Analyzed → ICP Refined → Next Campaign

Practical Usage Tips

1. Start with inbound lead qualification before outbound — inbound traffic is warmer, has lower brand risk, and gives you faster feedback to tune your qualification logic before you scale cold outreach
2. Never send cold email from your primary business domain — use 2-3 secondary domains with SPF/DKIM/DMARC configured and warm them for 3-6 weeks before launching
3. Multi-channel sequences (email + LinkedIn) deliver 287% higher reply rates than email alone — use at least two channels from the start
4. Set a maximum of 4-5 qualification questions before offering a meeting — over-qualifying a warm lead turns interest into annoyance and costs you deals
5. Keep bounce rates under 3% by verifying emails before each send; a damaged sender reputation takes months to repair
6. Pass every reply through an intent classifier before a human reads it — auto-answer FAQs, flag buying signals, and route positive replies within 2 hours during business hours
7. Let AI research, enrich, score, and draft — but keep human approval on customer-facing actions until you have validated quality over at least 100 sends
8. For capturing account research and company intelligence as persistent documents, see Personal Knowledge Base
9. For routing qualified leads through a structured sales pipeline, see CRM Sales Workflow