Strategy is Finally Changing
As AI changes strategy, Datent is pivoting and invites you on the journey
It is a truth universally sensed yet rarely admitted: corporate strategy is a broken tool-set, unfit for the modern organisation.
AI strategy should be the remedy, just as clever and disruptive as the underlying tech, but many “AI strategies” are as unimaginative and incremental as their corporate cousins.
Ask an exec team for their strategy and nine times out of ten you’ll receive a slick slide deck. Ask how it relates to customer needs and two more decks appear. Ask how it ties to the actual jobs that need doing and you’ll be bounced through two layers of translator-managers and five more slide stacks. The bureaucratic dance has become so familiar it is mistaken for leadership when, in practice, it is paper shuffling with origami level complexity.
Before we talk solutions, let’s briefly look at how we got here, and what strategy is supposed to achieve.
A very-brief long view
Stategy comes from the Greek word Strategos, meaning “leading the army.” The job was to survey terrain, weigh constraints and trade-offs, then cascade clear orders. This diagnosis, policy, action trio is still how strategy leaders like
teach strategy today.What has changed is the hierarchical cascade. As organisations scaled across geographies and disciplines, command-and-control reached its limits. No CEO could keep up with every context, so teams had to be empowered to devise and adapt local plans. Agile was born.
Agile’s rise was possible only because digital systems accelerated information flow (yes this is a nod to the blessed-and-cursed PowerPoint file). I once sat in a transformation workshop where a board member said, “The first thing we must consider in redesigning our operating model is that we are an SAP company.” Mission, customer, competitive advantage all took a back seat to the requirements of an Enterprise Resource Planning vendor.
Early IT was so rigid you had to bend the entire organisation around it. And yet, Rigid as they were, ERPs still beat paper ledgers. Each tech wave has boosted information flow and with it org agility.
AI offers the biggest step-change yet - but it needs us to rethink how we do strategy itself.
The cost of the deck-and-scrum combo
In short, we’re stuck in a dysfunctional middle ground … too rigid for real agility, too fragmented for coherent direction.
Command-and-Control / Waterfall
The Promise
Master plan keeps everyone aligned
Tight executive oversight
Clear accountability
Today’s Reality
Duplication everywhere - three marketing forecasts, four supply-chain dashboards for the same SKU
£100-300k decks as consultants explain the organisation to its own board
Outcomes forgotten as business cases required for investment are never audited and no lessons are learned
Empowered Teams / Agile
The Promise
Backlog regularly reprioritised as facts change
Cross-functional squads on one mission
Rapid learning kills bad ideas early a.k.a. fail fast
Today’s Reality
Time sink strategy decks are update quarterly (at best)
Silo’ed disciplines focus on furthering technical “maturity” scores before furthering their organisation’s mission
With a focus on outputs CDOs became Chief Dashboard Officers and CAIOs risk becoming Chief Agent Implementation Offices
It is tempting to blame agile, but agile rests on the very trade-offs great strategy requires.
Remember the Agile Manifesto? It won hearts by spelling out trade-offs and preferences like, ‘Individuals & interactions over processes & tools.’
Agile is not failing us, the issue is that technology hasn’t enabled a good enough flow of information to support scaled agile. The issue is slide decks and dashboards.
Why decks and dashboards can’t save us
Slides flatten nuance, dashboard flatten context. LLMs do neither.
Powerpoint can’t capture the nuances of strategy or the multi-dimensional nature of decent diagnoses. Dashboards improve information flow but are limited to quantitative signals that lack the context needed for decision making.
Large-language models change that equation. They scale qualitative analysis across transcripts, policies, Git commits (anything written), then pattern-match like a seasoned strategist and enable us to store strategy in a structured knowledge base that doesn’t dumb it down.
IF rich, nuanced signals can pass freely between executive and delivery layers, THEN alignment problems vanish. Teams know when another squad is tackling the same issue and execs can understand what's blocking delivery teams.
IF strategy lives in a modular knowledge graph, THEN it can be updated continuously. A customer requirement captured in a sales call today can be factored into a design and engineering planning session tomorrow.
No more PowerPoint handoffs up and down organisations and no more quarterly or annual efforts to overhaul strategy decks.
Enough theory - let’s get into practice, our findings and Datent’s pivot.
Where the penny dropped & Datent’s pivot
Datent began with a six-week Transformation Accelerator, our attempt to give change leaders the same lift that Y Combinator gives start-up founders. It was never meant to be the endgame, just a bridge-revenue experiment while we searched for a scalable way to make transformation easier.
Five cohorts later the programme holds a net promoter score above 90, kind alumni have called it “career-changing”, which is flattering as it has also been a huge source of learning for us.
Cohort 2 gave us our first AI jolt. A single prompt we wrote reproduced 7 of 10 recommendations a top-tier consultancy had delivered after a three-month, six-figure engagement. At that point it felt a little like a clever parlour trick.
There were many more AI proof points but the most powerful one for me came last December. After a sprint retro, our in-house AI coach flagged “founder’s newborn in NICU” as the top risk to our pivot roadmap. I’d only joined a few meetings and only mentioned the NICU commute only in passing at the start of the retro; none of us had named it as a risk. The model did and, of course, it was right to.
(Riley’s fine now, but that moment erased any lingering doubt about an LLM’s ability to surface what humans miss.)
Since then we’ve focused on putting our knowledge on data and AI strategy into a modular, pattern-based knowledge base. V1 had a dozen common data and AI strategies defined across over 50 dimensions and the results were better than expected:
A self-maintaining strategy library for members
Data-strategy report produced in hours and got instant board member buy in
Ability to produce transformation decks a client says would have cost £200k, generated at equal or better quality
All impressive but also at risk of missing the point.
The breakthrough is not slide automation not the ability to ‘automate consulting’ - an objective I find it baffling that so many startups have, we aim to work with them.
The breakthrough is the ability to hold strategy (with all its nuances on trade-offs, risks, and conflicting opinions) in a knowledge base that pairs perfectly with an LLM.
The earlier IF, THEN statements then seem very feasible and the implications point towards “strategy-as-code” and go way beyond the scope of this article.
Next - Strategy as Code and an AI Summer
After a long pause I’m excited to write here again with lots of articles planned on concepts such “strategy-as-code”, CI/CD strategy and much more1.
For Datent, we will be retiring the Accelerator2 after three final cohorts as we shift to the product focus we’ve hinted at throughout this piece.
If you’re curious about that there are demo sessions as part of a one offer AI Strategy Summer series we’re running as we lean into the impact of AI on strategy:
1. Open webinars (starting next week): Daily webinars next week, including two live Accelerator demos, and a follow-up webinar each month through the summer.
2. Summer Accelerator: First cohort with a free two-week trial, individual pricing, and a handful of scholarship seats for leaders currently between roles.
3. Four-week Virtual AI-Strategy Festival: Hands-on labs, leadership webinars, peer sessions all designed to fit around summer holidays. Because we are all still learning here and we believe that to be an innovative AI leader you need some hands-on experience. All for less than the price of a conference.
This is a one off programme as we shift to a product focus - register here for interest or more details.
We’re doing this now as we’ve always found the summer to be the best, if not the only, time to step back and reflect and we see that many people have an AI strategy itch to scratch.
Regular readers - thanks for waiting.
Toddler + newborn + startup killed my writing habit. I’m back with a ridiculous article backlog. I’d love feedback on the ideas in this one and to know what you’d like next.
Interested in acquiring the Accelerator IP?
We’re running it just three more times, polishing with every cohort (perfectionist in me can’t stop the iterations) and then shelving it to focus on the product. If you would value an applied leadership course with an NPS over 90 and rated as better than Berkley’s (from someone who’s been on both), let’s talk.