An AI voice agent that qualified leads for HP’s channel partners
A voice agent onboarded like a new hire, with a knowledge base and hardcoded compliance
Leads were flowing, sales weren’t
In commercial tech, over 90% of revenue flows through channel partners. These resellers are vital allies, but often under-resourced. HP had invested heavily in co-marketing. The leads were there, but too many slipped through the cracks without timely, qualified follow-up. Leads went cold and conversions stalled.
Windows 10 was nearing end-of-support, leaving 24M+ commercial devices ineligible for upgrade, and pandemic-era PCs were aging out. The window would not stay open. Partner interviews revealed that most leads were sitting dormant, and not for lack of interest. No one had time to qualify them. Partners didn’t need more names, they needed someone to verify contacts and route the right leads to sales. With BDR hiring off the table, HP built its own.
An AI agent built for outcomes
HP built an AI sales agent powered by Claude’s latest language models and integrated with workflow automations. It navigated IVR phone systems, confirmed contact details, asked buying questions such as whether a refresh was planned in the next 12 months, and flagged high-potential leads for human follow-up.
Signals layered over partner profiles
Partners supplied ideal customer profiles. An AI model was trained to look beyond firmographics, using signals like tech stack maturity and hiring velocity to identify similar accounts. The agent then ran outreach across those targets.
Onboarded like a junior BDR
Execution began with a clear, compliant conversation flow. Each call opened with AI disclosure and followed a logic tree built to guide the interaction. A central knowledge base held HP product information and partner context.
Safeguards written into every call
Objection handling was built in, from “not the right time” to budget constraints. Calls followed local time rules and attempt thresholds were capped, with DNC and TCPA compliance hardcoded. Human reviewers and AI simulations pressure-tested it before launch.
Sharper agents with every sprint
Two agent personas were tested for tone. “Stacey” was warm and conversational, “Scott” fast-paced and high-energy. IT decision-makers preferred Stacey. AI-powered review tools flagged dropout points and weak transitions, feeding insights back into the system each sprint.
AI earns its seat
The initiative reframed how partners think about sales enablement. With AI handling qualification, partner sales reps could focus on closing. The success triggered new pilots, including Sales Dojo.
Related work
The measure of a marketer
A controlled experiment across 500 accounts that measured what marketing added to deal size.
Lights, camera, avatars
Studio-quality AI avatars that replaced traditional video production and cut costs by 55%.
AI under the influence
A creator-led YouTube series that drove Copilot adoption without any paid media.