Running Companies in the Age of AI: From Rituals to Outcomes

I've had the privilege of working across small insurgents and large incumbents in TMT — hands-on turnarounds, platform builds, and board-level restructurings from Europe to the Gulf. We like to say people make the difference. True — but elite talent only compounds when three conditions are right: the attitude and profile of leadership, an operating framework that turns intent into action quickly, and a mindset that favors resilience, autonomy, and learning over theatre.

In the age of AI — where clarity, speed, and outcomes matter more than ever — failing to orchestrate these three dimensions leads to underperformance and, too often, irreversible decline.

In all the turnarounds I've done, the moment you realize a company is in trouble isn't when the P&L dips, but when the leadership team starts optimizing for how they look in the meeting, not what they do after it.

The Cost of Alignment Theatre

Five years ago, many believed Agile would unlock stiff organizations. Outside core IT, most transformations didn't change how decisions are taken nor returned their invested capital (e.g., PMI/Standish data consistently show high failure/re-baselining rates for large enterprise programs outside of core software delivery). Governance stayed the same, merely rebranded. Executive and steering committees multiplied; slide packs remained the centerpiece, designed to please egos and cover for poor delegation. Post-COVID, we kept the rituals on video and called it progress.

Internal meetings are cash burn. In many enterprises, senior leaders spend roughly four tenths of their time in internal meetings; managers report that a sizeable share is low-value. Companies as diverse as Haier, Amazon, and Buurtzorg operate in the 15-22% band: fewer status rituals, more real decisions. Our credible target under the new rhythm is ~20-25% for senior leaders and ~15-20% for managers.

It's time to rewire how we run companies so decisions move at the speed of reality, not the speed of a deck. Let's unpack the three dimensions — starting with the Framework that enforces the new behavior.

I. Framework — Retire Committees, Run the Business on Forums

Most companies default to periodic status: ExCos and SteerCos; curated metrics; approvals passed between forums. That cadence evolved for capital intensity and compliance, but it optimizes for alignment theatre instead of learning and speed. A lighter rhythm retires most committees and runs on a focused Framework.

A Caveat: Design to Fit, Don't Copy the Blueprint. This Framework is not a panacea. It's a set of design principles — a pattern for high-velocity decision-making. The forums, their cadences, and even the "Golden Metrics" must be re-engineered to fit your enterprise's physics and regulatory landscape. The non-negotiable principle is simply this: Retire the committees that optimize for status and replace them with focused forums that optimize for learning and speed.

The purpose of the Forums isn't the agenda; it's the behavioral compact they enforce.

Forums, Purpose, Behaviors

No Rank Hierarchy & Skin-in-the-Game: CEO / C-suite members may chair, but the people in the room are there because they hold data, own an outcome, bring options, or remove blockers. Attending meetings just to be "involved or informed" carries a clear opportunity cost.

Asynchronous Alignment: Everyone else stays aligned asynchronously — they watch the recording at double speed or read the summary. Better yet, they ask an AI assistant to summarize and cross-reference the decision with their current work, surfacing only what matters.

The Gravity Pull: Each forum lives on in its channel (e.g., Slack, Microsoft Teams). Teams feed inputs, debate trade-offs, and attach artifacts between sessions. Over time, the forums become gravity pulls — where problems, owners, and evidence naturally cluster — shortening cycles and raising decision quality and clock speed.

The Decision Log: Every bet and every decision from the forums is written to a public, immutable, searchable Decision Log (e.g., in Confluence/Notion/Dedicated Decision Log Software). No decision is official until it's in the log.

AI Powers the Truth: AI makes truth cheap by providing a single source of decision data based on golden metrics (Unit Margin, CSAT, Cost-per-Action). AI agents and specialized tools are instrumental:

• AI Meeting Assistants (e.g., Fireflies, Otter, Fellow) generate transcripts and summaries, which are searchable.
• Agent Pre-mortems (integrated with Jira, Asana, or ClickUp) scan historical project data, owner load, and dependency graphs to flag risks before the forum meets.
• Eval Bots (using BI/Model Governance systems and tools like Amplitude, Pendo) track adoption metrics and cross-link with financial and complaint trends to generate promotion/rollback recommendations.
• Generative tools (integrated into planning tools) propose option skeletons with impact/cost/risk ranges for High-Impact Solutions.
• Specialized AI Agents (built on internal data lakes or tools like Guru for knowledge management) combine unit economics with Customer/Data trends to auto-raise killcriteria flags.

The Framework does not replace every governance ritual, but it sharply limits their scope:

• Quarterly Business Reviews (QBRs): These remain, but shift their focus to synthesizing outcomes and trade-offs across the period, rather than being status reports.
• Monthly Architecture Discussions: These are preserved strictly for irreversible

choices only — platforms, data architecture, and major vendor commitments. The benefits are concrete: Decision latency falls. Reversals happen earlier (while they're still cheap). Budgets shift toward what works now, not what looked tidy in last quarter's deck.

II.

Mindset — Reversals as Professionalism

"Agile" promised speed; "simplification" promised relief. Outside IT and Engineering, both under-delivered because they didn't change the decision architecture. Incentives favored optics (green dashboards beat timely reversals). Technology debt made "simplify everywhere" naive.

A growth mindset isn't motivational wallpaper; it's amanagement stance: assume the system can improve if reality is visible and reversals are cheap. Two conditions make it real: psychological safety (bad news and half-built ideas surface without penalty) and doubleloop learning (goals and rules change when evidence demands it).

In the rewired enterprise, professionalism is defined not by winning every bet, but by terminating an underperforming bet before it crosses the kill-criteria threshold. Incentives shift from hiding bad news to acting on signal. Missing a kill criterion is a performance issue; meeting it is a career-accelerating decision.

III. Leadership — Same Courage, Now Wired Into Mechanisms

Great operators still set the intent, but their instincts must now translate into a repeatable mechanism that runs every week.

Instincts and Mechanisms

Instincts set the intent; mechanisms create the weekly behaviors that make truth cheap, decisions frequent, and reversals painless. 3 aspects are worth diving into: communication, delegation and dissent as they raise the bar for traditional status driven leaders.

The reduction in meeting time is not an excuse for silence; active communication remains the engine for mobilization and engagement. While the forums collapse status updates and cascades, leaders must strategically invest in how the results are shared. The Decision Log and asynchronous channels are critical for transparency, but leadership must actively use these tools to entice, mobilize, and drive action. The frequency and channel of communication must be contingent on the objective — not just informing, but translating the raw evidence and the Forums' weekly choices into a coherent narrative that middle management and the wider organization can act upon, ensuring they feel connected to the speed and progress of the core system.

The speed of the Framework is entirely dependent on explicit delegation of decision powers. It is not enough that the old ExCo no longer re-litigates; the owners in the new Forums must be formally empowered to decide within clear, published guardrails (budget caps, risk thresholds). This is the true meaning of the "skin-in-the-game" membership. If every decision from an Intent & Bets or Capital & Risk forum requires an upward review, the entire system collapses into the slow rituals it sought to replace. Leadership's primary act is granting this power, making it clear that accountability rests with the owner, not the committee.

A true learning system requires more than the courage to surface bad news; it requires the cultural safety to foster genuine debate and dissenting opinions. The forums are designed to bring competing truths together — functional conflicts, data disagreements, and philosophical differences on capability fit — and force them into a single, decisive action. This is inherently challenging to the authoritative leadership style still prevalent in many organizations. Leaders must actively model and reward constructive dissent, ensuring that opposing views are seen as critical inputs that sharpen the "Bet," rather than political resistance. Without this pressure testing, the system is prone to single-loop learning and confirmation bias.

The Transition: Assertive Cut-Over with a Safety Net

Switching from ExCos/SteerCos to the Forums creates a political vacuum. The path is an assertive, visible cut-over of the leadership agenda, coupled with a temporary technical safety net.

• The Hard Switch (Day 1): All senior leadership time previously allocated to status meetings must immediately be redirected to chairing/participating in the new Forums. The old ExCos and SteerCos remain on the calendar for 30 days — but are now designated as "Oversight Review" meetings. They receive the outputs of the Forums (the Decision Log) and can only veto a decision, not re-litigate it.
• The Proof-Point Loop: The Forums are given exclusive control over one visible, high-impact domain (e.g., all customer-impacting fixes).
• Retirement (Day 30): If the Forums demonstrate a faster cycle time or higher CSAT, the Oversight Review committee for that domain is permanently retired. This makes the transition proof-based and assertive, reducing the time for old rituals to regain political oxygen.

Conclusion

: Rewire First, Then Scale the Intelligence

This is not a tooling rollout. It is a leadership promise: truth will be easy to see, decisions will be easy to take, and reversals will be easy to do.

The Framework reduces the cost of running the company today. It also prepares the firm to welcome AI agents tomorrow. Agents need clean interfaces and humans who decide quickly. By 2027-2028, the highest-leverage move will be to give the strongest agents their own sandbox budget and a mini P&L. The agent that shows positive ROCE on its own actions graduates to larger mandates. Without this scaffolding, agents amplify chaos.

If you want a simple weekly test, make it this: Can we point to one decision we took faster, one reversal we made cheaper, and one ritual we stopped on purpose? That rhythm compounds. It's how you run the enterprise at far lower coordination cost while compounding effectiveness — leaner by design, smarter by habit, faster by choice.