Georgia-Pacific: Failed AI Sales Agent Led to Over 100 Deals
Manufacturer Georgia-Pacific's initial AI sales agent failed due to unreliability, but a redesigned version secured over 100 deals. The company now emphasizes a business-outcome-driven approach to AI implementation.

Manufacturer Georgia-Pacific has refined its approach to artificial intelligence implementation after an initial attempt at an AI sales agent proved unsuccessful. The first version was shelved because its responses were not reliable enough for employees to use.
The company's vice president of data and analytics, Matt Robuck, described the initial phase as a "technology-first approach," where businesses rushed to deploy AI agents without clear business objectives. This led to Georgia-Pacific's first agent failing to meet the company's standards.
Georgia-Pacific's second attempt involved starting with a defined business outcome and working backward. With improved data and context, the rebuilt agent was integrated into Salesforce and deployed to over 100 salespeople. Eight months later, Robuck reported that the agent had surfaced prospects leading to more than 100 deals.
Robuck stated that the experience demonstrated the significantly lower chance of success and business value when not starting with the desired outcome. This shift towards a business-driven strategy was critical for the agent's eventual success.