Why I’m ‘helping’ executive teams with agentic AI
Last month I shared what I am seeing in my work. It would appear that I am taking a less-than-charitable view of executives who have gone “all in” on agentic artificial intelligence (AI).[^ Going “all in” means ‘being fully committed, dedicated, or involved in a project, relationship, or decision—often without holding anything back. It signifies that someone is putting 100% of their effort, time, or resources into something and is willing to accept the risks associated with it.’ This is also a term used when folks are willing to gamble big; sliding their chips forward; “all or nothing.”] But I am nothing if not ruinously charitable. So, allow me this attempt to exonerate my narrative somewhat. I am, after all, a hedge wizard;[^ Yes, I know for a decade I called myself an “Archwizard of Ambiguity (most fantastic).” I still strive for ambiguity and fantasticness, but the archwizard bit always implied that I hold some sort of ceremonial role within The Academy. This isn’t actually true, and never really has been. I’m a rogue. And—like most of us lone practitioners who operate from outside of any lodge—my path to magic was self-taught via syncretic bricolage and cunning. And I’ve still much yet to learn.] I like to hedge, and keep my options open.
Yeah I do AI*
* Arcane Intelligence.
I work with executive teams who quest amidst the emergent. My talents are in what serves to develop in-house metis (practical wisdom and cunning) via participatory learning and the cultivation of fellowship and scenius (collective genius)—which in turn facilitates an attuned-acuity for the adjacent-possible. Ergo: strategic optionality.
Most of my work is—as I now like to say—“gestalt-informed.”[^ I say this because Kim—my darling partner Dr. Kim Lam aka the dangerlam, is almost half-way through a four year gestalt psychotherapy training program, and it vexes her lovingly to hear me speak so cavalierly about that which she is learning deeply. This is partly on account of my Mediocre White Man Powers, but largely because the overlaps between gestalt sensibilities, anthro-complexity and magic are significant.] There is a field to attend and attune to; a world of entangled pathways and shifting affordances. And AI—like it or loathe it—is most definitely within the field.
My contention, though, is that most organisations ought be questing amidst this field. To quest is to widen the lens of perception; to attend to the periphery and attune to the penumbra; to sense-into the edge of what we know, with acuity for anomaly.[^ Importantly, this includes that which we do not want to see. Things that might invalidate your hypothesis. Things that might even be ontologically destabilising, and inspire existential dread, hoho!] To quest is to follow the autopoiesis of your own enquiry, so as to find the fruits we might otherwise overlook (or never see).
But instead, when it comes to AI, I see much “decision-based evidence making.” A CEO—having spent far too much time in conversation with a large language model (whilst listening to investment propaganda from podcasts like All In)—decides that “AI is the Future!” And, following this, a strategy emerges wherein the organisational posture is to go “all in” on AI.
But then, months later, when the hype and honeymoon has faded, they are left to contend with the real complexities at play; along with a strategy that struggles to live up to its own hype.
Heh, just kidding, they don’t. It’s their direct reports that do. The executives charged and incentivised to execute the strategy.
Tensions are naturally high. Everything feels behind, and whatever prototypes have been demo’d fall vastly short of expectations. The workforce are fatigued; they’ve already been through so much change. “AI” was promised to make their life easier—but all it seems to be doing is reducing headcount (code for: making colleagues redundant) whilst also increasing workloads.[^ This is Jevon’s Paradox—a phenomenon I wrote of in the way back of How to Lead a Quest. The paradox is: as technological improvements increase the efficiency with which a resource is used, total consumption of that resource often increases rather than decreases. And so whilst AI “saves time”—the result is that we spend more time.] An air of desperation seeps in, and executives—they themselves entangled within the rivalrous dynamics of the socio-political context they call ‘work’—start to look for a magical new “solution” from outside of their context.
Often this translates to them hiring a new technology partner or consulting firm, which in many cases simply resets the same cycle.[^ But for a career executive you only need for the narrative to hold long enough to get some impressive preliminary results before you can “jump ship” into a new role.] But for those who want to escape this aspect of samsara[^ The endless cycle of illusion and suffering.] (or at least to side-step this aspect), we need to focus on conditions, not solutions.
This is what I—and many of my fellow liminal agents—do. This does not look like coming in with an answer, or a roadmap, or a fancy deck of explicit Thought Leadership to deploy.
Instead, we first attune.
Attunement
For a long time I have used words like attunement to describe a vital element of my work. But I’ve always done this in a vague and alluding manner, only gesturing towards what attunement means, or speaking of it indirectly, poetically.
I am very fond of words like acuity and attunement; meta-rational qualities shared by both wizard and rogue. And I’ve been meaning to write of this properly, so as to better teach it.
But then Dave Snowden himself recently wrote an excellent post on attunement—better than I ever could—presenting it as part of a trialectic: attune, orient and inscribe.
To attune is to bring yourself into right relationship with the system. It is the practitioner’s equivalent of the wood carver reading the grain: not yet deciding what to do, but developing sensitivity to what is actually present rather than what was expected. Attuning is never finished. It continues throughout any engagement, and the practitioner who stops attuning has stopped learning from the system, regardless of what else they are doing.
Attuning is not passive. It requires active, disciplined engagement: suspending the frameworks you arrived with long enough to notice what the system is actually doing, developing peripheral attention alongside focal attention, allowing surprise rather than filtering it. The practitioner who arrives with a methodology already running is not attuning. They are confirming.
We live in an age that lusts for confirmation. Confidence, certainty, clarity, conviction. The charismatic firm with the polished pitch and promised solutions.
But these approaches bring a pre-fabricated framework to impose upon a context that is not yet understood. They bring answers before having first attuned with the system itself.
Many executive teams also find themselves in interesting predicaments when their well-reasoned and very rational “plans” fail to account for nebulosity, entropy and the unknown. In an overview of meta-rational practices, David Chapman—another chap I greatly admire—shares a relatable example:
We’ll call situations in which rationality fails “anomalous”—in contrast with “typical” and “atypical.” Anomalous situations are ones that can’t be understood rationally. The software development plan was completely rational; it was replete with careful, empirically-based Excel-based cost estimates, personnel org charts, project GANTT charts, and system modularity diagrams. Reality started diverging from the plan almost immediately, for reasons no one could pinpoint. Eighteen months later, you’ve got an enormous incomprehensible mess. What happened? Rationally, everything should have worked! These are the kinds of situations in which meta-rationality may save the day.
Meta-rationality, as Chapman writes, is a vital competency for leadership. “Giant software development projects, with hundreds of millions of dollars invested, often fail due to missing or incompetent meta-rationality.” This is exactly what I am seeing with most leadership teams, when it comes to AI.
I also posit that an “acuity for anomaly”—the presence of mind to recognise when and where rationality fails, coupled with the wit to occupy ways of perceiving beyond rationality itself—shares much overlap with how one might approach “magic.” But that’s a tale for another time.
Back when I was a “king amongst the midwitted” I would have ready frameworks to deploy. Fancy models, and cantrips akin to “start with why.”[^ Face-validity is all you need.] These were neat tricks, very impressive. But I’ve long since become less enamoured with such sleight-of-mind. I particularly don’t like my work from a dozen years ago (before I wrote How to Lead a Quest), wherein I got swept up in ‘behavioural science’—a ‘discipline’ that seeks to reduce motivation and behaviour to rational systems. The result of such maneuver is that Life is rendered lifeless; we start to see things through the eyes of the basilisk—only that which can be abstracted, reduced, measured.
And so I now do my best to let all frameworks recede—including my assumptions—so that I can better attune to what is already at play. This is, as you know, part of the Trickster disposition.
“Better to operate with detachment, then; better to have a way but infuse it with a little humor; best, to have no way at all but to have instead the wit constantly to make one’s way anew from the materials at hand.” – Lewis Hyde, Trickster Makes This World: Mischief, Myth, and Art
Oh to serve whatever the situation asks of us! This is the work of a complexity practitioner, too.
But—this also makes for tricky positioning. What is it that we sell, exactly? What is the value we can point to, ahead of the time, before having attuned to context so as to discern what’s at play? And how do we articulate this value without this itself becoming its own attractor we then become biased to (thus influencing the field before we even intervene)? Also: are we just making it all up as we go? What’s ‘The Package,’ as it were? Or rather: what is the response to clients who ask “what are the outcomes we can expect?,” and say “let’s start with the end in mind, and work back from there”...?
Dave Snowden articulates the predicament of such an approach well in his Conditions, Not Conclusions essay:
The hardest thing about the Pelagian[^ Pelagian, Snowden writes, is where anthro-complexity practice operates. “The capacity for change is already present in the system. The practitioner’s role is to create conditions in which what is already possible becomes operative, to wayshape, not to supply. This does not mean doing nothing. Preparing the canvas, reading the grain of the wood, understanding the energy gradients, building containment strategies for unintended consequences: all of this is active, skilled work. It is simply not the work of supplying something the system lacks.”] mode is that it seems to imply less. It does not arrive with the confidence of a framework. It does not deliver a plan. It does not promise a transformation. It offers attunement, orientation, and inscription, and it waits to see what the system does with them. In a culture that measures practitioners by the ambition of their interventions, that is a difficult stance to hold.
But consider the alternative. The forlorn hope. The verloren hoop. The magnificent initiative that travelled fifteen hundred metres and sank.
I suspect that this is what we will see with regards to artificial intelligence: magnificent initiatives that end up costing so much more than any value they seemingly generate. Not in all cases—but in many.
The narrow path
So yes, I work with those who seek to harness the power of artificial intelligence. Which is akin to working with those who seek to use The One Ring for good; folks who—like Boromir—would say: it is a gift.
Which is to say: most (probably) mean well, but most are blind to power (and how it binds and corrupts) and consequence (unintended and n-th order effect). A wizard ought know the narrow path well; to work with power—without becoming corrupted by it.[^ A favourite related quote from Terry Pratchett’s Discworld: “Not doing any magic at all was the chief task of wizards—not ‘not doing magic’ because they couldn’t do magic, but not doing magic when they could do and didn’t. Any ignorant fool can fail to turn someone else into a frog. You have to be clever to refrain from doing it when you knew how easy it was. There were places in the world commemorating those times when wizards hadn't been quite as clever as that, and on many of them the grass would never grow again.”]
To walk the narrow path means to succumb to neither pessimism and doom, nor blind optimism and hope. It is also not to walk the mediocre middle, either. Instead, it is the ability to hold both perspectives at once—whilst remaining attuned to hidden ways yet to reveal themselves. The blue feather in the magpie’s tale.[^ “...to navigate mystery is not the same thing as living with uncertainty,” Martin Shaw writes. “It doesn’t contain the hallmarks of manic overconfidence or gnawing anxiety. It’s the blue feather in the magpie’s tale. Hard to glimpse without attention. There’s no franchise or hashtag attached. Navigating mystery humbles us, reminds us with every step that we don’t know everything, are not, in fact, the masters of all.”] The leaks.
To walk the path betwixt.
When I wrote How to Lead a Quest (over a decade ago), research in Artificial Intelligence was relatively open and academic. Some businesses I worked with were actively tracking development within their ‘quiver of options’ (a key part of the ‘scenario probing’ that informs Quest-Augmented Strategy). I used one of my client’s examples (a large law firm, with their permission) in the book. But this was all before the technological leap enabled by large language models.
I continued to watch with cautious enthusiasm, initially. But when Open AI went from being open to closed—my disposition shifted into deep caution. This is mostly on account of how power works, and the fact that AI is an accelerant to climate change,[^ “But AI uses less water than the beef industry!” some say. Yes, and I am glad you are concerned about water usage. But this is also a category error. AI functions as an accelerant within an already overshoot-bound system. It intensifies extraction rates, deepens supply chain complexity and fragility, makes coordination failures worse through race-to-the-bottom dynamics. Debating whether a prompt uses more or less water than a hamburger is a ploy that keeps you within a perspective-frame that needs to be transcended entirely.], ubiquitous corporate surveillance, and the metacrisis we share.[^ See my “Wisen up about AI—12 recommendations” from a couple of years ago; most of these still hold.]
Last year Ed Zitron wrote a piece titled The Era of The Business Idiot.[^ I am too nice to use words like “idiot,” and I don’t believe it is helpful to call any individual an idiot. But in aggregate we can appreciate how we have co-created the emergent situation where idiocy reigns. And I am okay with comedians and humorous writers “punching up.” It’s just that whenever I do it myself I usually regret it in hindsight. Anyway.] This essay articulates the backdrop against which AI is shaped.
[...] the problem runs a little deeper than the economy, which is a symptom of a bigger, virulent, and treatment-resistant plague that has infected the minds of those currently twigging at the levers of power — and really, the only levers that actually matter.
The incentives behind effectively everything we do have been broken by decades of neoliberal thinking, where the idea of a company — an entity created to do a thing in exchange for money —has been drained of all meaning beyond the continued domination and extraction of everything around it, focusing heavily on short-term gains and growth at all costs. In doing so, the definition of a “good business” has changed from one that makes good products at a fair price to a sustainable and loyal market, to one that can display the most stock price growth from quarter to quarter.
This is the Rot Economy, which is a useful description for how tech companies have voluntarily degraded their core products in order to placate shareholders, transforming useful — and sometimes beloved — services into a hollow shell of their former selves as a means of expressing growth. But it’s worth noting that this transformation isn’t constrained to the tech industry, nor was it a phenomena that occurred when the tech industry entered its current VC-fuelled, publicly-traded incarnation.
It’s not that “AI is bad”—the field of artificial intelligence is wondrous. Genuinely. But the field has become mostly captured, cordoned and corrupted by vested interests. And this is more precisely what many of us hold contention with.
Ed Zitron recently wrote a follow up to the essay above, titled Revenge of The Business Idiot. It remains polemic; giving voice to that which few have the courage to say. A sample:
LLMs impress the writers who do not want to write, the coders who don’t want to code, the researchers who don’t want to research, and the lawyers that don’t want to actually understand case law. Those that desperately tell you how powerful AI is and that you simply must use it are looking for you to validate their own laziness or distaste for effort, and those who are impressed with LLMs’ outputs tend to be people with low standards.
I don’t want folks to “feel bad” if they find themselves marvelling at the outputs of a large language model. I am one such person, at times. LLMs can be incredible—particularly if you don’t have subject matter expertise in the domain of your enquiry. But if you have accrued any metis in a particular field—any semblance of wit, wisdom and wiles—much of the output of LLMs remains dubious and underwhelming at best.
And yet: it’s here. And now two thirds of the client work I do comes from a context wherein there is a big commitment to integrate artificial intelligence. On briefing calls I’ll have senior leaders justify their strategic direction—with glazed yet confident eyes—“because AI is the future.” What do you mean by this? I want to say. But I need to tread carefully—for many have entwined their identity with the belief that “AI” is, indeed, “the future” [sic].
Eschew The Prophecy
Audrey Tang—former Minister of Digital Affairs for Taiwan—is one of my favourite people to listen to and learn from. I first heard her speak with Nate Hagens last year in this genuinely inspiring episode—Digital Democracy: Moving Beyond Big Tech to Save Open Societies. But recently she shared an apt post pertaining to the matter at hand: Reject AI Prophecies, Free the Future.
A growing number of organizations, from Silicon Valley tech outfits to a host of other big corporations, have begun linking AI use to performance evaluations. For many, survival in the workplace now involves investing substantial time and resources in learning how to collaborate with AI.
Yet, many may soon discover that these efforts do not necessarily make work smoother. Instead, they start to feel like obedience to a set of “prophecies.”
Consider claims such as: “AI will become your closest work partner,” “Everyone will have an AI assistant,” or “Companies that fail to adopt AI will be eliminated.” On face value, these statements seem to describe trends. In practice, they often discipline behavior: You better surrender your attention and judgment now, or you will be left behind.
My Oxford colleague Carissa Véliz, in her new book “Prophecy,” reminds us that the power of prophecy lies not in accurately predicting the future, but in shaping it.
For example, suppose a business owner believes that “AI will replace 80 percent of the workforce.” They may redesign performance systems, restructure teams and ultimately dismiss most of the employees. This does not prove that the prophecy was accurate. It proves that the prophecy is self-fulfilling.
The Prophecy is already shaping our vision of the future. What ought to be a dynamic constellation of visions that move us towards flourishing at higher orders of complexity (as per the gradient of the universe, Eros[^ See CosmoErotic Humanism: Philosophy in a Time Between Worlds By Dr. Marc Gafni & Dr. Zachary Stein.] and Life itself) is instead (temporarily) at risk of collapsing into what seems to be an inevitable techno-feudalist cyberpunk dystopia.
This future currently being sold to us by AI evangelists is a naïve vision of a future in which we have an abundance of time and energy and universal income; where AI agents do most of the work for us and our bigger challenge will be figuring out what to do with all the spare time we have. Except maybe this vision only applies to a small few—and there is now a rush to “escape the permanent underclass” (something Jasmine Sun expands upon). Folks are now scrambling, rolling the ladder up behind them, so as to make it to The Capitol and avoid life in The Outer Districts (aka what “the global south” already endure).
But... I don’t know. When things seem all-too-certain and sure (including in my own sense of things)—you know what this calls for.
The trickster disposition
For only that which can change can continue.[^ Thank you James P. Carse.]
Abi Awomosu, author of How Not To Use AI: 50 Contrarian Principles for the Imagination Age, writes of how we might learn from Trickster in how we relate to all of this. Her essay Your Imagination Was Always Empire’s Last Frontier is excellent, and articulates the positions we might take:
Once you see what is happening, the question is what to do about it.
The plays on offer are these. Each one carries an archetype — a position you can recognise yourself in, or recognise being asked to take.
Abstain. The Pure One. Refuse the tool. Stand outside. Keep your hands clean. The comfort here is moral purity. The cost is leaving the territory undefended in the moment it is being mapped. Absence in a time of enclosure is not neutral. It is a position. The empire counts on it.
Petition. The Reformer. Stay inside the system and ask it to change. Reform from inside. Plead for inclusion, representation, ethics review, alignment. The comfort here is the seat at the table. The cost is years of your real life poured into a structure that cannot give back what it is being asked for, and the slow atrophy of the muscle that knows how to do anything except ask. After enough years of petitioning, you can no longer imagine what taking would look like.
Surrender. The Convert. Don’t be left behind. Get ahead. Adopt the tool, be ready, you’ll be replaced by someone who knows how to use AI if you don’t move. The comfort here is pragmatism dressed as inevitability. The cost is the slow disappearance of the version of yourself that existed before the tool. You stop noticing what you used to be able to do. You stop missing the faculties that atrophied because you no longer remember they were yours. You become someone the empire could not lose, because there is no longer a you for them to lose.
Three plays. Three archetypes. All inside the empire’s frame. All running on comfort. Each with a cost the comfort hides.
There is a fourth play. It does not run on comfort. It runs on risk. Its archetype is the trickster.
To understand it, you need to understand the trickster.
The essay touches on similar themes that I have clumsily alluded to in the past; subversive leadership, heretical education, hiding in plain sight, surreptitious obliquity, feeding the leaks, hermetic disposition, and so on. The fourth play is the way.
I like the way Abi Awomosu thinks and writes. Here’s another excerpt:
Be read at the surface. Stay sovereign underneath. Empires need legibility to extract. Survivors learn to be legible at the surface and sovereign underneath.
That is the move I made with AI.
Most readers cannot place me. Am I pro-AI or anti-AI?
The question itself is the empire’s question — it forces a choice between Surrender and Pure One, the two plays that keep you inside the empire’s frame. I am neither.
The tension between legibility and sovereignty is a frustratingly generative one for me. Frustrating because I am so tired of this as a theme I keep seeming to write about. And yet, generative, too because: here I am, writing about it.
And so as you know, I do my best to obfuscate and baffle; to write honestly and non-linearly, as a form of “sacred self-sabotage.” At least, that is what I call it. But the real thing I am sabotaging is the trap(pings) of success.
I do not want to become so legible as to be rendered indistinct from the homogenised synthetica generated by large language models. At the same time, it is helpful in my profession as a speaker-facilitator and consultant to be at least somewhat legible, would you believe.
And so I keep coming back to this excerpt from my friend Tyson Yunkaporta from episode 321 of the Green Dreamer podcast.
I don't know, I try to avoid naming anything. And I try to avoid making too much sense, and I try to say things a bit differently every time and to mix it up. And I’ll make points that you can’t put together. I do that quite deliberately because I don't want the things I’m thinking or working on to become an ideology or a brand, or something that people can use as a name. I have seen that happen before with a few things I’ve done: People have grabbed it, and then it's become their thing. You’ve got to avoid that packaging and repackaging of ideas and let these things be free-range.
Tyson Yunkaporta deftly wields surface illegibility as a kind of cloak to protect the animacy/aliveness of what he is communicating; to ensure that what he shares isn’t subject to easy domestication and capture. On the surface, this is an inverse move to what Abi Awomosu writes of—she speaks of the cloak of legibility as a survival strategy—but the effect is the same. The sovereignty, imagination and aliveness are protected.
I’m not sure what the move is, for me. I daresay it will be to oscillate. To stay in motion, and not overcommit to any particular register.
Sometimes I worry: will I lose my warm construct-aware meta-rational mythopoetic post-achievement metamodern-adjacent trans-perspectival complexity-savvy collapse-aware readers if I were to slip into highly legible so-called “Thought Leadership”? And will I lose my client base whenever I wax lyrical—loquaciously and without any comforting linearity—in a manner not easily reduced nor conveniently rendered into pithy and practical “tips”?
It’s the perennial question. I guess we shall see.
Towards a world more curious and kind
And a future less grim.
This is what I keep coming back to; the beacon I orient towards.
This is why I’m ‘helping’ executive teams with agentic AI. And I’m well situated to do so (provided I don’t blow my cover with too much talk of equality, freedom, and a liveable planet).
Most folk in tech are good people who—like many of us—‘fell’ into their role. They don’t want to be bought by Palantir, nor build the Torment Nexus. They want to use technology to help people. They are rational people navigating situations in which rationality alone will not suffice. And they’re doing the best they can with the resources they’ve got. As is the case for many of us.
Life does not afford most of us the vantage from which we can sense how our choices fit within a wider trajectory. And so we rarely perceive how power laws, rivalrous dynamics and perverse incentives shape the conditions within which we play.
And so the pressures these executives face are palpable. The board wants a strategy, the market wants a signal, and the workforce wants certainty. To “go all in on AI” was, at the time, a seemingly reasonable move. Of course, if they had adopted the Quest-Augmented approach to strategy I write of in How to Lead a Quest, they wouldn’t have leapt to drastic “all-in” strategies. They would have been cultivating “a quiver of options,” which is an implicit practice of attunement, orientation and inscription (via experimentation). This makes strategy less of a gamble or a punt, and more of an organic, emergent and participatory knowingness.
The folks I work with are wise enough to know we are on a narrow and treacherous path—yet still they believe there is a future in which businesses make good products at fair prices to a sustainable and loyal market. A future in which organisations can improve the quality of the value they generate. Even if they are cursed with knowledge of the metacrisis—some still carry a tenderness despite. The glint of a fellow infinite player, liminal agent, and kindred spirit.
We all have a part to play in this unfurling pantomime, this infinite game we share. And who knows—we may yet surprise ourselves. Let us see.
So yes, I’m ‘helping’ executive teams with agentic AI, he says (legibly). But really, I’m helping teams to stay human; alive, awake and attuned to the opportunities at hand. Walking the narrow path so as to find ‘the fourth way’ in which we might play to continue the play.
I am also working towards teaching the wit, wisdom and wiles of being rogue wizard, complexity practitioner and liminal agent. The School of Fox Wizardry will be coming (soon!™)—stay subscribed to learn more.
Member discussion