Building Lucy: An AI Demand-Generation Platform for B2B Teams

The Challenge

B2B demand-generation teams juggle a messy stack: one tool to define an ideal customer profile (ICP), another to build audiences, a third for campaigns, and a spreadsheet to tie it all together. Lucy was built to collapse that stack into a single AI-native workspace — from "who should we target?" all the way to an activated pipeline of buyers.

What We Built

AI-Generated Ideal Customer Profiles

At the center of Lucy is an ICP generator. A user describes their motion in plain English — or points Lucy at their website — and the platform proposes a structured ICP: target industries, company sizes, geographies, and the decision-maker roles that matter. Marketers refine it instead of starting from a blank page.

Audiences, Campaigns, Buyers and Insights

Around the ICP, Lucy is organized into four workspaces:

**Audiences** — turn an ICP into a concrete, de-duplicated list of accounts and contacts.

**Campaigns** — launch full-funnel programs against those audiences and track them over time.

**Buyers** — the people behind the accounts, enriched and scored.

**Insights** — live results: buyer counts, accounts reached, contacts per account, and geographic spread.

A Home That Remembers Context

The dashboard greets users with recent work — the audiences, campaigns, and ICPs they touched last — so they can pick up where they left off rather than re-navigating the whole tool.

The Stack

**Next.js + React** for a fast, app-like front end

**A dedicated enrichment pipeline** for resolving and scoring accounts and buyers

**LLM-backed ICP generation** grounded in the customer's own site and inputs

**Multi-tenant data model** so each workspace stays isolated and auditable

Results

Lucy replaces several point tools with one guided flow. Teams go from a blank ICP to a saved, activated audience in minutes, and keep a single source of truth for every campaign they run — exactly the kind of AI product we love building at Runtym.