Microsft’s CEO posted an article in X yesterday about how the real value of AI for organizations is not in what the AI model can do for you, but in the learning loops you can architect between your knowledge, the AI you use, and back again. Awesome! But so much more is possible.
Much of the press conversation is focused on the second part of his post, about how there has to be an ecosystem of technology providers and not just a few big models capturing all the value and destroying whole industries.
That’s where he stopped short. And the next step seems both obvious and really exciting to me. Was his oversight accidental, or intentional? I don’t know.
Let’s start with the first part, because they are very connected – or they should be.
Learning loops with your own knowledge
Nadella: “This means the real opportunity is not in picking the best model but instead in building a learning loop on top of models where human capital and token capital compound. You can offload a task, or even a job, but you can never offload your learning. The future of the firm is the ability to compound that learning across people and AI….Companies need to turn their workflows, domain knowledge, and accumulated judgment into AI systems that improve with each use…This loop becomes the new IP of the firm.”
So far, this is great. And it’s an idea whose time has come. A lot of people are thinking and building that way, from RAG to MCP to various B2B systems that let you chat with, or have things analyzed in light of, your existing custom knowledge. As an individual or organization. For example, Luke Wroblewski wrote a nice post last month about what he calls Collaborative Steering. (“Today, AI tools are mostly solo sports. Developers write more code. Designers create more images. PMs crank out more docs. That’s cool… but it’s more cool to work together. So how does that work with AI?”)
Myself, I usually don’t chat just with Claude directly, I chat with a Claude Project populated by my Obsidian reading and thinking notes that I’ve taken over the last 10 years.
As a result, it says things like “Marshall, I’m not your therapist, but you’ve shared enough notes from your therapist with me that I can say this: your problem couldn’t be more clear!” Oh, Claude.
But more generally, you could say it helps retrieve relevant things I’ve written down over the years, from the cold storage parts of my memory, back up into my short term working memory, with amazing speed and thoroughness. This incentivizes me to take better notes, because I know they’ll not just come in handy someday maybe if I remember, but now they will be put in my hands at just the right time in the future.
The loop: There’s a lot of my favorite insights from Claude in those conversations that gets written back down in my notes. It’s then referenced again later in other conversations.
I think this is the kind of architected learning loop Satya Nadella says is the key form of value for companies in using AI. And as a consultant, I do deliver extra value by being able to more effectively utilize everything I’ve written down in years past.
Why stop there???
But then Nadella goes on to say something really striking, from a business perspective, while stopping short of the lion’s share of the opportunity here!
He says: by the way, we need a whole ecosystem of AI providers that you can swap out as part of your human-token learning loop. Not just one or two capturing all the value; no, there needs to be an ecosystem of platforms generating even more value on top of themselves than they capture with their own businesses.
Fine, but if we are going to talk about how the real value is in learning loops with prior knowledge, and about how there’s got to be a thriving ecosystem, let’s take this to the logical conclusion, let’s close the loop, and say that all our organizations operate in ecosystems with various degrees of thriving and we can build and capture incredible value and impact by building a learning loop with AI and not just our own knowledge– but the collective knowledge of the whole ecosystem!
In the social media era, there was a saying: the smartest person in the room is the room. At its best, social media is not just a new star-maker with another winner-take-most economy, just delivered on new platforms. At its best, social media is a collective expression, discovery, and sense-making force for positive collective impact!
Even if you don’t buy that, though, why would you stop with a feedback loop based on your own company’s knowledge?
Every day, new networks of people, organizations, and machines, all around the world, find new insights based on analysis of their knowledge in aggregate.
From the mundane (UK power transmission provider SSEN Transmission joins the European Network for Cyber Security as a Knowledge and Information Partner) to the strategic (Theory of Change as a shared learning practice).
So why not architect learning loops between AI, your organization’s knowledge, and the work of other organizations all around you?
Systematic market monitoring, structured analysis, automated determination of what could constitute net new knowledge – so much is possible, in fact, it’s right there in front of us.
I was pretty shocked to read the CEO of Microsoft stop at (a) the real value of AI is in learning loops with your knowledge, and (b) we need thriving technology ecosystems, without closing the loop with (c) we should build learning loops with our ecosystems.
But maybe that’s his next article. Or maybe that’s what will hit the mainstream in a few years.
I believe in the future it will be absolute business malpractice to fail to pay attention, systematically, to knowledge and activities outside your own walls.
I’ve been doing it for years and thats what I’m building with organizations now.