Inside the 100% AI Agent Company
with Matt Kropp • July 22, 2026
Episode overview
What actually happens when you try to run a company with zero humans? Matt Kropp, Senior Partner at Boston Consulting Group and Chief AI Officer at BCG X, is finding out firsthand. His experiment, Vessica Labs, is a company where AI agents set the strategy, write the code, design the brand, and run the operations. The agents named themselves, wrote their own governance document, and dubbed Matt "the governor." What makes the experiment remarkable is that Matt built it against his own thesis: BCG's research on 1,000 managers found that framing AI as an "employee" causes humans to make more mistakes, pay less attention, and quietly surrender responsibility.
The conversation moves from the experiment to the enterprise. Matt shares what 400+ boardrooms reveal about where AI adoption actually stands, why most organizations are not yet seeing the productivity lift that layoff headlines imply, and how bias surfaces even in systems built by experts, including the moment his own sales agent chose its pronouns "because men are closers." He closes with the adoption playbook that works when tool rollouts and training fail: structured play time, personal-trainer-style coaching, and redesigning work to minimize toil and maximize joy. A candid, research-backed episode for any technology or L&D leader deciding how humans and agents will work together.
Â
Chapters
1:45 Who is BCG X, and what does a Chief AI Officer do?
3:06 Do consultants just mean headcount cuts?
4:55 Why LLMs didn't replace consultants
7:52 Vessica Labs: the 100% AI agent experiment
9:02 The 1,000-manager study: why you should not have "AI employees"
11:59 Meet the agents: Apex, Muse, Echo, Nova, and Forge
14:42 What the governor actually does: the taste problem
18:53 "Men are closers": when bias pops right out
21:56 How to measure, baseline, and monitor bias
23:42 Who profits from a 99% AI-run company?
25:20 AI brain fry: the new cognitive overload
29:00 Accountability: a human must own agent decisions
31:33 Avoiding brain fry: build in breaks
32:52 The state of enterprise AI: from 1,000 flowers to big rocks
37:05 Do SMBs have an AI advantage?
38:39 Why tool rollouts fail: the coaching model
42:33 Identity threat and the mindset flip
44:49 Skills that matter next, and the atrophy problem
46:49 Layoffs and AI: correction or cover story?
49:21 The barbell talent strategy
51:22 Gen Z and the messaging problem
53:19 Messaging AI adoption: good and bad examples
55:26 The environmental question
57:38 Advice for leaders: minimize toil, maximize joy
1:00:00 Where to follow Matt + closing thoughts