The Hansen First Brain and the McKinsey Second Brain: A Short Q&A

Posted on November 20, 2026

0


Both learn from outcomes. The difference is what the loop does when the result falls short.


With increasing interest in what we do, one of the more common questions we are asked is this: what is the difference between the Hansen first brain and the McKinsey second brain, and how do the two work together? Here are the short answers.

One point first, because it frames everything that follows. The difference between a 1998 agent and a 2026 agent is that one is human and the other is AI. In 1998, the agents were people: technicians, buyers, suppliers, couriers and customs brokers. Today, AI agents work alongside those same people, and both are part of the same equation. The first brain has to account for every agent in the operation, human and AI, because each one acts on what it is measured on and what it can see.

What is McKinsey’s “second brain”?

In October, McKinsey described an AI operating system that works as an enterprise “second brain.” It maintains the company’s context, reasons across functions, recommends or executes decisions, and learns from outcomes. Its newest graphic puts that learning in the foreground: decisions, actions and outcomes feed back into the system to improve what it recommends next.

What is the Hansen first brain?

It is the organization’s validated understanding of how it actually operates, not how it is documented. It is built by starting with the outcome being produced today and tracing backward through the people, incentives, handoffs, external agents and workarounds producing it. That understanding is what any AI operating system should be reasoning on.

Isn’t learning from outcomes new?

No. In 1998, RAM fed every transaction back into the next decision. Each supplier’s standing was updated from what happened before the order, during delivery and after it, combining historical performance with real-time conditions. People weighted price against delivery and added what no system could capture yet, such as weather affecting a delivery area. All of it ran inside firm fences: defined limits on what the system could do, when it had to stop or redirect, and which human was accountable.

The loop was one part of the result. Tracing back to how the operation actually ran, changing the practices behind the shortfall, and then running that loop together took next-day delivery for the parts operation serving Canada’s Department of National Defence from 51 percent to 97.3 percent in three months, and held it there for seven years.

So what is the difference between the first brain and the second brain?

McKinsey’s second brain learns from outcomes within the context it has been given: the codified processes, policies and institutional knowledge of the enterprise. The Hansen first brain asks whether that context is true before anything learns from it.

McKinsey’s own case study tests agents against business outcomes to know whether the context has been fully captured. That is the right test for capture. The first brain asks the question before it: whether what was captured is true.

The difference isn’t whether a system learns from outcomes; both do. It’s what the loop does when the result falls short. The second brain refines its next recommendation and extends the context it has codified. The first brain reopens the operating picture, traces back to the strands the record never captured, and redraws the fences around what it finds.

One learns within the map. The other keeps checking whether the map is right.

Why does that matter more with AI?

Because an AI agent is still an agent. In 1998, technicians measured on call volume held their parts orders until 4 p.m., even though a policy told them to order immediately. The guardrail existed. The incentive won. An AI agent given a target will optimize for that target the same way, only faster and at greater scale, and it sees only what it is connected to. Agents need firm fences drawn around how the operation actually runs, whether the agent is a person or a model.

Do the two brains compete, or work together?

They work together. The first brain feeds whichever second brain an organization adopts, today’s or the next one.

How the two operating systems fit together. On the left, the operation runs as parallel strands. Phase 0™ traces backward from the outcome and brings to light the strands the documented map leaves out, within a firm fence of named human authority, and real outcomes loop back into it. That validated operating reality feeds McKinsey’s AI OS on the right. Illustrative.

McKinsey’s own first rule is “Own the context; rent or replace the technology.” The context worth owning is the one that has been validated against how the operation actually runs.

We have been running that loop, on the record, since 1998.

I wrote about how the two operating systems fit together here: How the Hansen-McKinsey AI Operating Systems Address Choudary’s “Human-in-the-Loop Fallacy”

-30-

Coming next: the other second brains

McKinsey is not the only firm building an enterprise second brain. Over the coming weeks, we will take the same questions to three more:

Two-by-two positioning chart titled "Where the brains sit: technology capability vs. grounding in operating reality." The horizontal axis is technology capability and scale. The vertical axis is grounding in operating reality, meaning whether a system validates its context against outcomes and traces back past the documented process. The Hansen first brain (ARA RAM 2025 on SLAP OS) sits in the upper left, labeled "Grounded, built to feed a platform." A solid arrow labeled "validated operating reality" runs from it to the upper right quadrant, "Grounded and capable: first brain feeding second brain." A dashed arrow returns, labeled "Infinity Loopback: actual outcomes come back and reopen the operating picture." Four second brains sit in the lower right quadrant, "Capable at scale, grounded in the documented process": McKinsey AI OS (EcliptOS) as a solid dot, assessed in this post, and Accenture AI Foundation plus AI Factory, Globant Glob.AI OS, and Deloitte plus Palantir EOS as hollow dots, with assessment to come. A side panel explains that the second brains learn from outcomes inside the context they codify, while the first brain checks whether that context is right. A box reads "Roots in 1998: method formulated 1998 (SR&ED); DND next-day delivery 51% to 97.3% in three months, held seven years; named the first brain in 2026." The footer offers Pilot Rescue as a first step and states that positions are assessed on what is publicly available today and updated as research continues. Disclosure: Hansen Models offers the first brain.

Where the brains sit. McKinsey’s second brain is assessed in this post. The other three are placed from the design each describes in its own public materials, and each will be assessed in its own Q&A. Each assessment rests on what is publicly available today, and we will update it as our research continues. Disclosure: Hansen Models™ offers the first brain.

  • Deloitte and Palantir: the Enterprise Operating System (EOS). It combines Deloitte’s domain experience and AI agents with Palantir’s Foundry and AIP platforms, with Palantir’s ontology pre-configured using Deloitte’s leading practices.
  • Accenture: AI Foundation and AI Factory. Announced on 21 October 2026, it combines NVIDIA’s AI infrastructure with Accenture’s workflow orchestration, governance and industry intelligence as a governed foundation for enterprise agentic AI.
  • Globant: Glob.AI OS. Launched on 27 October 2026 as a unified layer for governing and adopting AI across the enterprise, across more than 320 AI models.

Each will get the same short Q&A: what the second brain is, what context it learns from, what its loop does when a result falls short, and how the Hansen first brain works with it.

Posted in: Commentary