The pyramid cracks. What agentic AI does to the consulting leverage model.

OpenAI’s February 23, 2026, partnerships with BCG, McKinsey & Company, Accenture and Capgemini have turned agentic AI from a consulting topic into a direct test of the industry’s own delivery model, because the same systems being sold to clients can automate work long handled by junior consultants.

OpenAI said its Frontier Alliances will pair its product and engineering teams with major consulting partners to help companies define strategy, connect systems, redesign workflows and deploy AI agents across organizations. The named partners are among the firms most associated with the traditional consulting leverage model: senior partners sell and steer work, while larger teams of junior staff perform research, analysis, modeling and slide production.

The current news development is not that the consulting pyramid has disappeared. It is that leading firms are now formally tied to the rollout of enterprise AI agents, increasing the pressure to prove whether large project teams remain the right unit of delivery. In a September 2025 Harvard Business Review article, David S. Duncan, Tyler Anderson and Jeffrey Saviano argued that AI is pushing consulting firms toward a narrower structure in which fewer junior staff support senior expertise with more automated work.

Thorsten Meyer AI framed the issue as a crack in the pyramid, asking what agentic AI does to the consulting leverage model. That framing is an interpretation, not a confirmed outcome. What is confirmed is the growth of agent-based enterprise AI programs and the participation of major consulting firms in taking them to market.

Why It Matters

The consulting pyramid matters because it has shaped how firms make money, train talent and price work. The model depends on leverage: relatively expensive senior partners and managers supported by a broad base of lower-cost analysts and associates. If AI agents can perform parts of research, data preparation, draft analysis, project tracking and presentation work, clients may question billing structures tied mainly to human time.

For readers inside consulting, the near-term issue is career structure. Junior roles have historically served as both production capacity and training ground. If firms hire fewer entry-level consultants or give them different work, the industry will need new paths for developing judgment, client skills and domain expertise. For clients, the issue is whether AI-enabled delivery should reduce cost, speed projects, change contract terms or shift fees toward outcomes.

The change also matters to AI vendors. OpenAI’s move shows that enterprise AI adoption is not only a software sale; it requires workflow redesign, data access, controls, change management and integration into existing systems. That is why consulting firms remain central even as AI challenges parts of their staffing model.

Background

The consulting leverage model was built around a pyramid: a small number of partners at the top, a middle layer of managers and senior consultants, and a wide base of junior staff. That base handled much of the repeatable work that supported high-margin advisory projects. Generative AI first put pressure on tasks such as summarization, research and slide drafting. Agentic AI extends that pressure by linking models to tools, data sources and multi-step workflows.

OpenAI’s Frontier Alliances announcement said BCG, McKinsey, Accenture and Capgemini would help customers put AI coworkers to work across organizations. Separately, the Harvard Business Review article described a shift from a broad pyramid toward a leaner model as AI handles more work previously done by junior consultants.

“AI coworkers”

— OpenAI

“define strategy, integrate systems, redesign workflows, and scale deployment globally”

— OpenAI

“The pyramid cracks.”

— Thorsten Meyer AI

What Remains Unclear

It is not yet clear how much agentic AI will reduce consulting headcount, change promotion paths or lower client fees. Public announcements confirm partnerships and investment in AI delivery, but they do not prove that the pyramid model has already been replaced. Outcomes will depend on client adoption, reliability of AI agents, data controls, regulation, liability rules and whether firms pass productivity gains to clients or keep them as margin.

What’s Next

The next phase will be measured in project economics rather than announcements. Watch for changes in entry-level hiring, consultant utilization, project staffing ratios, fixed-fee or outcome-based pricing, and disclosures from major firms about AI-driven productivity. Clients are also likely to ask for clearer separation between human advisory work and automated delivery.

Source: Thorsten Meyer AI

Key Questions

What actually happened?

OpenAI announced multi-year Frontier Alliances with BCG, McKinsey & Company, Accenture and Capgemini on February 23, 2026. The partnerships are meant to help enterprises deploy AI agents in real workflows.

Does this mean consulting firms will stop hiring juniors?

No confirmed evidence shows a full stop. The more limited confirmed point is that AI can take on some work often assigned to junior consultants, which may change hiring volumes, role design and training.

Why is the pyramid model under pressure?

The pyramid depends on a broad base of junior staff doing repeatable research, analysis and production work. AI agents may perform parts of that work faster, which challenges staffing and billing assumptions.

Will clients pay less for consulting?

That remains unclear. AI may reduce delivery costs, but firms could also use those gains to improve margins, bundle proprietary tools or shift pricing toward outcomes.

Source: Thorsten Meyer AI