-> back to projects

Case Study

public case study

Agentic outbound system for enterprise GTM

An AI-powered outbound workflow built for Transcend to identify higher-intent accounts, enrich and qualify them, generate account-specific messaging, and route next steps into the sales workflow.

STAR framing

Situation

Enterprise outbound needed an automated, on-demand framework for turning account signals into qualified research, messaging, and rep-ready workflows.

Task

Create a repeatable system that could identify high-intent accounts, suppress low-fit outreach, and route better context into sales workflows.

Action

Built Clay workflows and Claude reasoning layers for enrichment, scoring, account research, outbound angle generation, and CRM handoff.

Result

Contributed to $500K+ in pipeline per quarter while reducing manual research and improving the relevance of outbound motion.

Context

Enterprise outbound teams have to turn scattered market signals into timely, relevant outreach. The work often spans account research, enrichment, qualification, message generation, CRM updates, and rep handoff. I built a system to make that process more repeatable, data-driven, and easier to operate at scale.

Problem

  • Manual research made it hard to prioritize the right accounts at the right time.
  • Signals lived across disconnected tools and were difficult to compare consistently.
  • Outbound quality depended on how much context a rep could gather before writing.
  • Low-fit accounts needed suppression logic so automation did not create noise.

Architecture

The pipeline starts by detecting which signals actually matter — matching raw events to defined buying triggers and validating ICP fit. Those signals are then ranked for target personas so the right contacts surface first. Clay handles enrichment, scoring, and suppression; Claude drafts three fully customized emails per contact, tailored to persona and signal context. Qualified contacts enroll in an Outreach sequence with that custom content loaded — not a generic template — and reps review before launch.

Example transformation

Sanitized account signal in, qualified outbound package out.

input

account: Fortune 500 retailer
signal: AI initiative announced (earnings call)
contact: VP Digital Platform
persona: platform_leader
icp_score: 87
trigger_match: ai_readiness
persona_rank: 1
suppress: false

output

angle: AI readiness blocked by data permissions
priority: tier_a
persona: platform_leader
sequence: enterprise_outbound
emails:
  - ai data readiness at {company}
  - why permissions block ai launches
  - {company} + transcend next step
enrollment: outreach · 3 custom steps
status: ready_for_rep_review

What I built

  • Defined ICP, buying-trigger, and suppression logic for account prioritization.
  • Built signal detection and persona-ranking layers to match triggers and prioritize contacts.
  • Built Clay workflows for account sourcing, enrichment, scoring, and routing.
  • Designed Claude prompts for research synthesis, qualification, and a 3-email outbound sequence per contact.
  • Created validation steps to keep enriched fields reliable across CRM workflows.
  • Integrated workflow signals with Salesforce, Outreach, Gong, Slack, and data providers.

Example workflow

  1. A target account shows a relevant market signal, such as hiring, funding, or GTM activity.
  2. The system detects the buying trigger, ranks contacts by persona fit, and checks suppression rules.
  3. Clay enriches the account; Claude picks a persona-specific angle and drafts three fully customized emails.
  4. The contact enrolls in an Outreach sequence with all three email steps pre-written for that account.
  5. Reps review the sequence content, then launch outreach with structured CRM metadata attached.

Impact

The system contributed to $500K+ in pipeline per quarter by improving account prioritization, reducing manual research time, and making outbound messaging more context-aware.

Public-safe note

This case study intentionally uses a sanitized workflow description. It excludes internal screenshots, customer data, proprietary prompts, CRM fields, and confidential operational details.