The Boston Consulting Group Isn’t Wrong

That doesn’t make them right.

Every system you own makes a fundamental assumption about you: that you’re predictable. The question your organisation never asks is whether the next generation should make the same assumption.

That’s the difference between Systems of Record and Systems of Engagement. Same systems. Same technology. Different thinking about what they’re for.

For two years I’ve been writing about this distinction. It’s not a technology story. It’s a thinking story.

Same systems. Same technology.
Different thinking about what they’re for.

This is what Human By Design is about. Not building new systems but rather changing how we think about the ones we’ve already built, designing for the future and making space for people to be unpredictable, to surprise, to lead.

Even today, organisations choose the first path. They record history. They extrapolate. They execute. And that is rarely questioned because people don’t realise that there is a choice.

In the ‘Age of Engagement’ choosing wrong is catastrophic.

The Misunderstanding

Mention the need for ‘Systems of Engagement’ and people hear ‘CRM’, ‘triggered emails’, ‘personalisation engines’ et al - and assume they’re covered.

They are not.

CRM is a loaded euphemism. It sounds like relationship management because that is the pitch. The fact is your CRM is fundamentally for ‘Customer Records Management’ - it records data.

The system ‘knows’ you, but it (and the organisation) ‘doesn’t understand you’.

And we never stop to think about it. Muscle memory, autopilot, sleepwalking - whatever you call it - our thinking and questioning goes out the window. Context gets lost.

The system knows you. The organisation doesn’t understand you.

What they’re actually describing (though they use the language of engagement) are ‘Systems of Record’. Still based on history. Still optimised for prediction. Still assuming the next action is predetermined by the last one.

The thinking didn’t shift. Only the deployment. The governance model is identical. The control architecture is identical. They’ve just repackaged the same operating model in shinier language.

Boston Consulting Group

This weekend a friend sent me Part 3 of an ongoing BCG series: ‘The Science Behind Next-Best Action Programs’¹

It was new to me, but given who sent it, I read it. Turns out this is the third of three papers, with two more to come. It was very clear why they thought I needed to read this.

BCG are describing a different approach to how systems work. Coming at it from architecture and technology. The infrastructure required to support what they call the ‘agentic era’. They’re right about the problem.

BCG is primarily describing the path to technical implementation. The real work seems to be buried.

Which means they’re missing the real gap. You can have the most elegant three-layer stack in the world. If your organisation can’t articulate what it actually believes, if your leadership hasn’t made the operating model explicit, if you haven’t aligned on what responsiveness means. Then you’ve just built a very fast machine that amplifies confusion at scale.

You don’t need AI to amplify confusion. You need thinking to cure it.

One Problem. Two Lenses.

Both are describing Systems of Engagement. BCG describes the technical architecture, the systems. PHI⑊PIN describes the underlying philosophy that makes the architecture worth building.

BCG’s three-layer stack (propensity scoring, contextual bandits, foundation model agents) only works if the thinking underneath it is sound. If you don’t know what your organisation actually stands for, if you haven’t made your operating model explicit, if you haven’t aligned on what responsiveness means, then the most elegant architecture amplifies confusion at scale.

BCG gives you the how. PHI⑊PIN starts with the thinking that determines whether the how actually serves your organisation.

This is why organisations burn through AI budgets and see nothing. Action without thinking has been the problem for decades. You don’t need AI to amplify it; you need thinking to cure it. The smartest companies in the world are discovering this the hard way. They build capability without clarity. They deploy without conviction. Then they wonder why the system works beautifully but the organisation is confused.

This isn’t new. It’s as old as business itself. But now you can do it at scale, at speed, with the ‘confidence’ that the problem is the tool and not the thinking.

BCG has published three of five papers. Two more are coming and I think that this is where the rubber will meet the road. The architecture is elegant. The ‘science works’. BUT - architecture without coherence is just very fast confusion.

Let me not attempt to predict. When I read them I will return with part two.

Why Humans Break Every Pattern

Jacob Collier made a point about why humans are better than AI.

The best people are unpredictable. They catch the zeitgeist precisely because they don’t follow a script. Not even their own.

The Population Problem

Systems of Record assume predictability. They build models on historical behaviour and extrapolate. They’re fundamentally designed to assume tomorrow looks like yesterday. I won’t get into the emerging ‘prediction market category’ but in many ways, it’s doubling down on the SOR approach. Humans are assumed to be predictable.

If you read Asimov’s Foundation Trilogy - pay close attention to Hari Seldon’s core insight. Psychohistory works. It’s elegant, powerful, mathematically sound. But it only works at the population level. It predicts the behaviour of masses, trends, civilisations. The moment an individual acts from conviction rather than pattern. The moment someone surprises. The entire prediction collapses.

That’s the gap between Systems of Record and Systems of Engagement. SOR assumes the system works on individuals the way psychohistory works on populations. Predictable inputs produce predetermined outputs. Scale the pattern. Automate the response.

But humans aren’t populations. They’re individuals. And the ones who matter most are the ones willing to surprise.

Systems of Engagement assume unpredictability. They don’t predict what humans will do. They create conditions for humans to do surprising things and learn from it.

Systems of Record assume you’re predictable.
Systems of Engagement assume unpredictability.

---

The playing field doesn’t level because we built a bigger machine. It levels because we rethought the problem.

BCG will publish two more papers. I’ll read them. And I expect the architecture will get more elegant still. But elegance is not the bottleneck. Clarity and conviction are. The willingness to ask what your organisation actually believes before you ask what your system should do next.

That’s the work. It was always the work.

The conversations are happening. The commitments are real. If you’re sensing this and already asking these questions, then you don’t need a better system - yet. You need a clearer answer to a simpler one: what are you actually trying to do?

That’s where we start.

1. Silvio Palumbo, Yun Lim, Nabeel Siddiqi, Mark Abraham, Karl Johnson.
May/June 2026 - Boston Consulting Group.