Cranked to 11: Superagility and Alignment
A deep dive into how organizations must balance enhanced AI-driven agility with strategic alignment in the era of rapid transformation
If an organization hopes to master transformation as a repeatable skill and achieve sustainable relevance for the long term, it eventually has to learn to balance agility with alignment. This is one of the great tensions in modern leadership – and as we’ve argued here before, one that absolutely must be tackled with an AND approach.
In an operating environment of rapid change and high uncertainty, is agility – defined broadly here as the ability to experiment and iteratively learn and unlearn at speed – critically important? Yes.
And in that same operating environment of rapid change and high uncertainty, is alignment – defined here as the ability to share a purpose, act on a vision, and keep the main thing the main thing – also critically important? Yes.
Now, which one is more important? Also yes. One isn’t sufficient without the other. In fact, one doesn’t even make sense without the other.
The ability to rapidly experiment, prototype, and learn doesn’t hold much value unless it’s directed toward a valuable end. Similarly, a great learning goal – e.g., a great question – or a shared vision or purpose isn’t worth much to an organization’s future if teams don't have the ability to push from the problem space into the solution space through agile experimentation and learning.
So far, so… this-is-all-pretty-well-understood, right?
Well, what if we’re in the process of redefining the scale of agility in a growing set of domains? And instead of a knob that goes from, say, 1 - 10, we’re not only cranking it to 11 but actually twisting that agility knob up an order of magnitude?
In our Briefing last Friday, Pascal linked to an essay from Geoffrey Huntley that sketches a picture of what I’d call “superagility” in software development. In a companion post on X, Huntley summed it up with:
“Ya know that old saying ideas are cheap and execution is everything? Well it's being flipped on it's [sic] head by AI. Execution is cheap and ideas are now expensive.”
At least in software development, the data looks likely to back up this argument about execution. The AI-enabled development platform Cursor recently hit $100 million ARR faster than any company in history. And the just-published Anthropic Economic Index shows that software engineering is the top current use case for Claude (representing ~37.2% of user queries).
As my friend Michell Zappa puts it:
“[C]oding now feels less like a privileged domain and more like a straightforward way to strengthen our personal agency. We no longer need to be professional developers to give shape to our ideas… When you realize you can whip up a working prototype in days (or even hours), your perspective on what’s possible broadens exponentially…”
Now, if we look to the history of agile learning as a guide, perhaps we can expect superagility to follow a similar path of being proven out in software development before jumping to a much broader set of domains as a transformative approach and even mindset.
I think this is entirely plausible and even probable. Sam Altman gestured toward something like this (again) in his latest blog post exploring a world with ubiquitous intelligent agents:
“[I]magine it as a real-but-relatively-junior virtual coworker. Now imagine 1,000 of them. Or 1 million of them. Now imagine such agents in every field of knowledge work.”
Even if we fall well short of every field, the prospect should raise some urgent questions. Specifically, how do we coordinate learning when experimentation and iteration can move exponentially faster than the processes we’re used to managing? How do we maximize the value of superagility?
I think the answers will likely bring us back to where we started – with the idea that the value of agility critically depends on appropriately valuing alignment. And in a world of superagility, the value of building and maintaining alignment is only going to increase. As organizations (rightly) rush to enhance their abilities to experiment and iterate through adoption of new AI tools, they’d be wise to spend commensurate time thinking about how those enhanced abilities fit within a purpose-aligned learning system – what we’ve described before as a high-performing symphony rather than a room full of super-empowered self-directed soloists.
So, how are you thinking about superalignment in the age of superagility?
@Jeffrey
I like how you rightly point out the agility/alignment form of Barry Johnson's polarity map here: unsolvable rather binary problems that contextually depend on continual re-weighting between the two.
But agents in every field of knowledge work? This definition and envisioning of superagility is heavily burdened with reductionist thinking. The idea that everyone in a company is maximally efficient by ruthlessly pursuing their own individual goals without ever being burdened to slow down and coordinate/recoordinate with each other any.
This is also one of the contributors to CEOs now questioning the limits of company effectiveness - not so much efficiency - when no one physically meets anymore (in a shared office or otherwise).
What's missing is organizational intelligence is not the same as individual intelligence. UC Berkeley ML/AI legend, Michael I Jordan, touched on this last week at the AI Action Summit in Paris by bringing up the microeconomics, collective intelligence, and mitigation of uncertainty by markets that these reductionist visions of AI agents do not possess:
https://www.youtube.com/live/W0QLq4qEmKg?t=3811s
Love the concept of superagility and totally agree with the 'AND' of superagillity ALIGNED with your organization's purpose, vision, and values. Four years ago Pascal highlighted that Finance and Accounting are operating in an environment of rapid change and high uncertainty and that finance needs a new skill - agility – defined broadly here as the ability to experiment and iteratively learn and unlearn at speed. This just adds a big exclamation point to this with the concept of 'superagility'. Great work and thanks for continuing to inspire us Jeffrey & Pascal!