BCG educates AI sales agent Jamie on its top-selling products as well as its biggest errors.

      **TL;DR** Boston Consulting Group is developing an AI sales agent named Jamie that learns from both the successful and unsuccessful sales behaviors by analyzing call transcripts and engagement patterns. This approach aims to coach human salespeople and evolve over time. Vercel has adopted a similar strategy, reducing its sales development representative team from ten to just one human managing an AI agent. In 2025, BCG reported $3.6 billion in AI consulting revenue, constituting 25% of its total revenue of $14.4 billion, and is part of OpenAI’s Frontier Alliances for enterprise AI deployment.

      Boston Consulting Group is taking a somewhat unconventional approach with its new AI sales agent by intentionally training it to recognize what doesn’t work. Named Jamie, the agent learns from the call transcripts, engagement techniques, and conversational styles of both BCG’s top sellers and those whose methods were less effective, aiming to understand not only successful tactics but also to avoid ineffective strategies. Over the past two years, the industry has primarily focused on the capabilities of AI, but BCG is now emphasizing the importance of understanding what should be avoided.

      Japjit Ghai, managing director and partner at BCG X, the company’s tech division, elaborated on this method in a recent episode of BCG’s podcast, *The So What from BCG*. “We trained the agent by analyzing the top sellers, their call transcripts, and customer interactions, while also ensuring Jamie doesn't mirror the worst seller experiences,” Ghai explained.

      Regarding Jamie's learning process, Ghai offered further insights in an interview with Business Insider. Jamie utilizes various sources: BCG’s internal knowledge, client insights, and existing sales calls and transcripts. Ghai noted that these recordings constitute a “repository of often underleveraged assets” that BCG can analyze to identify effective behaviors. The intent is not to duplicate the actions of any specific seller but to uncover historical trends that indicate which behaviors resonate with customers and which do not.

      Additionally, the system is intended to provide coaching for the human sales representatives it partners with. Following a call, a seller receives a tailored scorecard from Jamie that highlights their strengths and weaknesses based on their conversational techniques. Each interaction generates data that continuously enhances the model, making Jamie what Ghai describes as an “always-on muscle” that perpetually improves.

      BCG is careful in its messaging, framing Jamie not as a replacement for its sales team, but as a distillation of accumulated institutional knowledge designed to make the instinctive practices of the firm’s best performers accessible to all. However, it remains to be seen if this approach holds in the face of AI deployment economics; a system that successfully replicates a top seller's behavior could inherently create a need for fewer salespeople, regardless of the original intent.

      The dynamics have already manifested in other companies. For instance, Vercel, a developer cloud platform, followed this principle by training an AI agent based on its best-performing sales development representative, leading them to reduce their SDR team from ten to just one human overseeing the AI. The remaining nine employees transitioned to outbound prospecting roles instead of being laid off, signaling the clear message that once an AI agent learns what the top employee does, fewer salespeople are necessary.

      David Totten, Vercel’s vice president of global field engineering, remarked in October that “modeling after top-performing employees has always been standard practice. The difference now is that technology enables us to accelerate this process.” The company shared that the AI agent now handles inbound qualification, filters spam, and reroutes leads—tasks that previously required the entire team.

      For BCG, Jamie represents a significant strategic investment in AI. The firm reported revenues of $14.4 billion for 2025, with approximately $3.6 billion (25%) coming from AI-related consulting work, according to a disclosure first highlighted by Bloomberg in April. AI and technological services now comprise over 40% of BCG’s total revenue. The firm has expanded its workforce to include 33,500 employees, bringing on AI engineers, data scientists, and IT architects alongside traditional consulting teams.

      BCG X, where Ghai works, was formed in late 2022 by consolidating three existing subsidiaries—BCG Digital Ventures, BCG Platinion, and BCG Gamma—into a unified unit of over 3,000 technologists, scientists, and designers. This division creates AI-driven solutions for clients such as L’Oreal, BMW, and New York Life. Thus, Jamie serves not only as a tool for BCG's internal operations but also as a demonstration of the capabilities the firm markets: the idea that AI agents can learn from an organization’s top performers and operationalize that knowledge on a larger scale.

      The broader consulting industry is also moving in this direction. OpenAI formed partnerships with BCG, McKinsey, Accenture, and Capgemini in February to assist in the sale and implementation of its Frontier AI