Name: Lia Golledge Company: MagicBlocks Function: Growth
Date: September 2024 – May 2026

TL;DR

This work spans three connected roles.

• As a consultant and strategist, Lia worked directly with a customer — understanding their sales problems, their lead flow, and their conversion gaps — and helped them build the strategy behind their MagicBlocks AI agent.
• As an operator, she maintained and optimised MagicBlocks' own live AI agents (originally built by the CEO): updating the knowledge base, refining the playbook, improving goal logic, and connecting the agents to HubSpot so qualified leads flowed directly into the sales pipeline.
• And as a builder, she created a library of inbound and outbound demo agents across multiple regulated industries — mortgage, insurance, finance, home services — giving the sales team credible, ready-to-show proof for every prospect conversation. The result was a +60% increase in contact conversions and a conversion infrastructure that now underpins both the PLG and sales-led motions.

KEY OUTCOMES

+60% Increase in contact conversions from optimised AI agents

781 Users acquired through growth engine including agent-driven capture

+938% Website traffic growth in 6 months (agents as conversion layer)

Multi-vertical Demo agents built across mortgage, insurance, finance & home services

Live HubSpot workflows, pipelines, lifecycle automation & enterprise deal flow running

Building MagicBlocks AI Agent

THE DESIGN THINKING JOURNEY

1. Empathise — understanding the problem

Three different groups had three different problems. The customer Lia worked with had inbound leads but no consistent way to engage them quickly — the gap between lead arriving and rep responding was costing them conversions. They knew they needed an AI agent but didn't know how to configure one that would actually work for their sales motion.

MagicBlocks itself had a different problem. The CEO had built a working inbound agent, but the knowledge base needed ongoing maintenance, the playbook needed refinement as the product evolved, and the HubSpot connection wasn't set up to route leads cleanly into the sales pipeline. Good enough to be live. Not good enough to be a proof point.

For the sales department, the problem was credibility. Prospects in mortgage, insurance, and finance would ask: "Do you have something that works in my industry?" Without a live demo agent built for their vertical, the answer was too abstract. A library of working, industry-specific agents would change that conversation entirely.

2. Define — framing the right problem

For the customer engagement, the real question wasn't "which agent features do they need?" It was: what does their lead flow actually look like, and where does it break down? Lia approached the conversation as a diagnostic — mapping their current motion before recommending a configuration. The output was a strategy, not just a setup guide.

For the MagicBlocks agents, the framing shifted from "keep this running" to "make this a conversion engine." That meant treating knowledge management, goal logic, and HubSpot mapping as optimisation levers, not maintenance tasks. Each one had a measurable effect on whether a lead became a deal.

For the demo agents, the problem was specificity. A generic "insurance agent" wouldn't be convincing. The frame became: build agents that reflect the real objections, qualification questions, and conversion goals of each vertical — so a prospect sees something that mirrors their world, not a generic demo.

3. Ideate — Exploring the Possibilities