Matt Blumberg, CEO of Markup AI, has pioneered an innovative approach to executive decision-making by creating a “fantasy board of directors” powered by artificial intelligence. The four-time technology CEO built this AI agent to supplement his actual human board, seeking diverse perspectives beyond the limitations of his five real board members.
The creation process began with a fantasy draft-style selection where Blumberg and his executive team compiled a spreadsheet of prominent figures from various categories: iconic business leaders, tech CEOs, venture capitalists, authors, thought leaders, and diverse voices. The final roster of approximately 15 members includes legendary figures like Warren Buffett, Steve Jobs, Oprah Winfrey, and Taylor Swift, alongside people Blumberg knows personally, including VC Fred Wilson, with whom he’s worked for 25 years.
The technical implementation involved AI generating 5,000-word profiles for each fantasy board member, using templates designed to capture how these individuals would react to business situations based on their actual public statements and known philosophies. The team then loaded these profiles into an AI agent, along with comprehensive instruction sets about the company’s operations, past board meeting materials, quarterly business review decks, and other relevant documentation. The entire setup took just one to two hours.
Blumberg uses this AI board as a strategic thought partner, typically consulting it every other week. Applications include reviewing board materials before sending them to his real board, identifying potential questions or gaps in presentations, developing themes for company meetings, and even receiving performance reviews. Recently, the fantasy board delivered a 2025 performance review that Blumberg described as remarkably accurate, correctly identifying both his strengths and weaknesses—a assessment his executive team found “pretty impressive.”
However, Blumberg acknowledges two significant limitations: First, the AI only knows what it’s been told or what’s publicly available, lacking the day-to-day context of company operations. Second, it cannot replicate the human elements of real board interactions—body language, facial expressions, and the subtle cues that come from in-person strategic discussions. He emphasizes the importance of not blindly trusting AI outputs, noting that “they’re predictive, and they’re good, but not perfect.”
When Blumberg presented this innovation to his actual board in November, they responded enthusiastically, with several members requesting roadmaps to build their own versions. The real board viewed the fantasy board exactly as intended: a valuable augmentation of thought partnership, never a substitute for genuine human advisors.
Key Quotes
We have an actual human board, and it’s fantastic. I get really good advice from them. But I wanted to use AI to create a fantasy board because mine is limited by the five people who are on it and the experiences that they have in life.
Matt Blumberg explains his motivation for creating the AI-powered fantasy board. This quote establishes that the AI tool is designed to augment, not replace, human advisors—a critical distinction in responsible AI implementation for executive decision-making.
One thing we put in the instruction set is that when I ask the board for its opinion, very generically, I want it to tell me the consensus opinion and notable outlying dissent with quotes of things that the fantasy board members have actually said.
Blumberg describes the sophisticated prompting strategy used to make the AI board more realistic and useful. By requesting both consensus and dissenting views grounded in actual statements, the system provides nuanced perspectives rather than generic advice.
With any AI agents, you need to be really careful not to believe the bullshit. They’re predictive, and they’re good, but not perfect. They’re not human, and they miss a lot of cues.
The CEO offers a candid warning about AI limitations, demonstrating responsible leadership in AI adoption. This acknowledgment is particularly important given his role leading an AI company, showing he understands both the technology’s potential and its boundaries.
They viewed the fantasy board the same way I do, which is that it’s a nice add-on, a sort of augmentation of thought partnership, but that it’s never going to be a substitute for a real board.
Blumberg describes his real board’s reaction to learning about the fantasy board. Their enthusiasm and interest in replicating the approach, combined with shared understanding of its limitations, validates the concept while maintaining appropriate boundaries for AI use in governance.
Our Take
Blumberg’s fantasy board represents a pragmatic middle ground in the ongoing debate about AI’s role in business leadership. Rather than the extremes of AI replacing executives or being relegated to mundane tasks, this demonstrates AI as a cognitive diversity tool—expanding the range of perspectives available without eliminating human judgment.
What’s particularly noteworthy is the grounding mechanism: using actual quotes and documented philosophies rather than allowing the AI to fabricate responses. This approach addresses one of the most significant risks in AI-powered advisory systems—hallucination and unfounded recommendations.
The quick implementation time is deceptive. While the technical setup took hours, the real value comes from thoughtful curation—selecting which voices to include, what materials to provide, and how to structure queries. This suggests that effective AI deployment in strategic contexts requires as much human wisdom in design as it does technical capability.
Blumberg’s transparency about limitations, especially given his position leading an AI company, provides a refreshing counterpoint to AI hype. His framework—AI as augmentation, not replacement—offers a replicable model for executives exploring AI’s strategic applications.
Why This Matters
This story represents a significant evolution in how AI agents are being deployed for executive decision-making and strategic planning. Rather than replacing human advisors, Blumberg demonstrates AI’s potential as a complementary tool that expands the range of perspectives available to leaders. This approach addresses a common limitation in corporate governance—boards are constrained by the specific experiences and viewpoints of their members.
The implications extend beyond individual companies. As AI agents become more sophisticated, executives across industries could leverage similar systems to stress-test ideas, anticipate stakeholder reactions, and explore diverse strategic options before committing resources. The relatively quick implementation time (1-2 hours) suggests this isn’t just a concept for tech companies with massive AI budgets, but potentially accessible to a broader range of organizations.
This use case also highlights the maturation of AI from task automation to strategic partnership. By creating detailed profiles based on real statements and philosophies, the system moves beyond generic responses to provide contextually relevant advice that mirrors how actual thought leaders might respond. However, Blumberg’s candid acknowledgment of limitations—particularly around context and human nuance—provides an important framework for responsible AI adoption in high-stakes decision-making environments.
Related Stories
- CEOs Express Insecurity About AI Strategy and Implementation
- Microsoft AI CEO’s Career Advice for Young People in the AI Era
- IBM CEO’s Vision on AI’s Future and Business Impact
- PwC Hosts ‘Prompting Parties’ to Train Employees on AI Usage
- Business Leaders Share Top 3 AI Workforce Predictions for 2025
Source: https://www.businessinsider.com/ceo-used-ai-to-build-a-fantasy-board-of-directors-2026-1