Home Impact Perspectives About

Cobalt Perspectives

Practitioner thinking on AI and commercial value creation.

A five-part series for commercial leaders, operating partners, and transformation executives who need to move from understanding AI in principle to deploying it in practice.

The series

Most organisations have deployed AI tools. Fewer than 20% are capturing measurable EBITDA impact from them. The gap is not technology — it is operating model. This series builds the case for why, and then provides the practical frameworks to close it.

Parts I and II make the case. Parts III, IV, and V are the work — architecture, diagnostic, and execution. Each article stands alone, but the series is designed to build. Articles are available on request — use the link below each article to request your copy directly.

The Series

AI and the Operating Model Gap

Part I · The Case

AI and the Operating Model Gap

Why organisations are failing to capture the AI opportunity — and a practitioner's framework for closing the gap from diligence to exit.

Most PE-backed businesses have an AI strategy. Fewer than 20% show measurable EBITDA impact. The gap is not the technology — it is the operating model: the decision rights, incentive structures, and accountability frameworks that determine whether AI tools produce commercial behaviour or just dashboards. This article identifies five failure modes and the structural fixes that separate the funds capturing AI value from those accumulating AI spend.

"Delivering 20% IRR now requires 10–12% annual EBITDA growth — driven almost entirely by operations. AI is not optional in this environment."
Request this article →
The Multi-Agent Operating System

Part II · The Case

The Multi-Agent Operating System

Why the next wave of AI value creation has nothing to do with prompts — and everything to do with architecture, orchestration, and agent-to-agent commerce.

74% of companies plan agentic AI deployment within two years. Only 9% currently orchestrate agents across workflows. This article maps the architecture inflection point — from isolated tools to coordinated operating systems — and explains why the orchestration layer is simultaneously the most important investment and the least understood. Includes governance design principles that separate scalable systems from expensive pilots.

"Orchestrated systems deliver 2.5× the productivity of isolated tools. 55% of agentic AI capital is now invested in orchestration — not applications."
Request this article →
The Commercial AI Stack

Part III · The Work

The Commercial AI Stack

What it is built from, how the layers connect, which tools sit in each — and what it looks like deployed across pricing, sales, customer success, and operations.

Three layers: Data & Integration (the foundation), Orchestration (the value multiplier), and Commercial Applications (where impact lands). Most organisations deploy application-layer tools on a broken foundation — which is why pilots succeed and scaling fails. This article maps the full tool landscape and provides six documented workflow transformations with measured impact ranges, from dynamic pricing (+4–6pp gross margin) to AI-native customer service (60–80% Tier 1 resolution).

"80% of organisations cite data limitations as the primary scaling roadblock. The orchestration layer is what most organisations are missing entirely."
Request this article →
The AI Value Diagnostic

Part IV · The Work

The AI Value Diagnostic

A scored self-assessment framework for identifying where your organisation actually stands — and where to focus first.

Before deploying the next tool, answer six questions: Is the data layer clean enough to produce reliable outputs? Does a commercial owner with a specific EBITDA number own the outcome? Are AI outputs embedded in the decisions that drive commercial behaviour? This article provides a scored diagnostic across six dimensions — producing an Era 1/2/3 positioning and a prioritised action list specific to your commercial model and business archetype.

"A plan without ownership is a wish list. The diagnostic identifies not just the technical gaps — but the accountability gaps."
Request this article →
The 100-Day Commercial AI Programme

Part V · The Work

The 100-Day Commercial AI Programme

A structured execution framework for moving from diagnostic to measurable commercial impact — for PE-backed businesses and corporate transformation leaders.

95% of organisations have an AI strategy. Fewer than 6% are capturing more than 5% of EBIT from it. The gap is not budget — it is structured execution with commercial accountability. This article provides the complete 100-day programme: three phases (Diagnose → Sprint & Build → Embed), the five failure modes that kill programmes before and after Day 100, and archetype-specific starting points for transactional B2B, complex enterprise, and recurring managed services businesses.

"100 days is long enough to demonstrate real commercial impact and short enough to maintain organisational focus."
Request this article →

On LinkedIn

Shorter-form thinking on PE value creation, transformation, and AI deployment.

Follow along on LinkedIn for regular posts on operating model design, value creation, and the practical realities of AI deployment in PE-backed businesses.

Follow on LinkedIn →

If you are working through a growth challenge, a value creation plan, or an AI deployment programme — let's talk.