AI only works when your system is ready for it

Don’t see AI as a feature. See it as an infrastructure.

Most companies are experimenting with AI.

Few are redesigning how revenue actually works.

That distinction matters.

AI does not create advantage because it writes emails faster or scores leads more precisely. It creates advantage when it is embedded into the structural logic of how revenue is generated, protected and expanded.

In other words, AI becomes valuable when it moves from tool to system layer.

For complex B2B organisations, this is not optional. It is architectural.

Start with the commercial outcomes, not the technology

An AI-enabled commercial model does not begin with a platform decision. It begins with clarity on three revenue priorities:

1. Protect revenue already earned, by actively monitoring account health, renewal risk and customer satisfaction, ensuring existing revenue does not quietly erode.

2. Expand revenue within existing accounts, by identifying cross-sell and upsell opportunities aligned with real customer needs and long-term value creation.

3. Acquire new revenue with precision, by targeting the right accounts, prioritising high-intent opportunities and focusing resources where win probability is strongest.

These are board-level concerns (and not just marketing objectives).

AI contributes when it strengthens decision quality across these three areas. That means earlier visibility into churn risk, clearer identification of expansion readiness, and sharper prioritisation of new accounts.

But to do that reliably, AI must sit on top of a coherent system.

Let’s explore.

The four structural layers of an AI-enabled commercial model

Think of the model as four connected layers (as in, remove one and the system weakens).

1. Unified revenue data

Most organisations do not suffer from lack of data. They suffer from fragmentation.

AI is an enabler, an amplifyer. And it will amplify whatever it is fed.

And if the data foundation is inconsistent, the outputs will be inconsistent too.

Now, a unified revenue layer means:

• Shared lifecycle definitions across markets, with everyone speaking the same language when it comes to leads, opportunities and customers, so performance means the same thing in every region.

• Harmonised account and contact structures, structured around real buying groups rather than internal silos, allowing enterprise relationships to be managed as one instead of fragmented by market.

• Integrated commercial and behavioural data, connecting sales activity, marketing engagement and customer signals to create one coherent view of account health and intent.

• Clear ownership of data governance, with explicit responsibility for definitions, data quality and system logic, ensuring the infrastructure remains aligned with strategy over time.

To be clear, without this, AI does not scale. It fragments.

More on this: Your CRM isn’t just software, it’s your commercial infrastructure.

2. Clear lifecycle signals

Data alone does not guide action. Signals do.

Signals? By signal, I refer to a meaningful behavioural deviation or acceleration within a defined lifecycle stage.

For example:

• A strategic account whose buying rhythm slows compared to its segment norm

• A customer whose usage pattern mirrors those who historically upgraded

• A new domain showing clustered research behaviour across multiple stakeholders

In short, signals connect behaviour to decision-making.

They answer a practical leadership question: Where should we intervene right now to see the biggest impact?

When signals are clearly defined, AI can continuously monitor patterns across thousands of accounts and surface what no team could manually detect.

This can move an organisation from being reactive to being anticipatory.

3. Shared metrics and decision logic

AI really becomes powerful when marketing, sales and customer success act on the same interpretation of signals.

Of course, that requires shared metrics.

If marketing optimises for lead volume while sales optimises for late-stage conversion, the system fragments. If regions define “qualified opportunity” differently, AI outputs lose comparability.

For instance, an efficient AI-enabled operating model should (at least) align around:

• Revenue contribution (not activity volume)

• Lifecycle progression (not isolated campaign results)

• Account-level performance (not siloed channel metrics)

This is where alignment and operating model design intersect with AI.

Technology cannot compensate for organisational misalignment. It can only accelerate it.

When positioning and lifecycle ownership drift across regions, scalability falters long before technology can fix it.

4. Action automation and human escalation

Of course, another reason to implement AI in your operations is to transform insights into concrete and impactful actions. And so:

• once a churn signal crosses a threshold, something should happen.

• once expansion likelihood rises, something should activate.

• once early buying energy appears in the market, outreach should adjust.

That may mean:

• Automated nurture sequences, guiding prospects with relevant content over time, so engagement continues even when sales is not directly involved.

• Task creation for account managers, triggering timely follow-ups based on buyer behaviour or lifecycle stage, reducing reliance on manual tracking.

• Dynamic prioritisation of outbound queues, continuously ranking accounts by intent and potential, so teams focus on the opportunities most likely to convert.

• Or adjusted budget allocation across markets, reallocating investment based on real performance and pipeline signals rather than fixed annual assumptions.

Disclaimer: regardless of how neat your AI implementation is, human judgement should remain critical. Use AI to identify probabilities and flag trends, but always rely on humans to handle nuance.

Governance is what makes AI scalable

What we observe from our daily interactions is that B2B organisations often struggle not with ambition, but with variation.

• Different regions interpret positioning differently.

• Different business units define value differently.

• Different teams adopt tools differently.

An AI-enabled commercial model therefore requires governance at three levels:

1. Data governance: ownership, definitions, quality standards

2. Signal governance: agreement on what constitutes churn risk, expansion readiness, buying intent

3. Commercial governance: clarity on how actions are triggered and measured

Without governance, AI insights remain local experiments. Governance is what turns AI from experimentation into enterprise capability.

This is where you form your competitive advantage (not in isolated AI pilots, but in structurally embedded intelligence).

From experimentation to revenue infrastructure

Experimentation is key. But it cannot be your end game.

Playing with pilots, scoring models or chatbots on your website is useful, but it does not yet redesign the commercial infrastructure.

An AI-enabled commercial model looks different.

It connects:

Strategy + Positioning + Lifecycle management + Data structure + Signals + Actions + Revenue operations.

The entire chain is deliberate.

What’s more, the impact compounds. Early churn detection improves retention forecasting. Better expansion signals increase customer lifetime value. More precise acquisition reduces wasted spend.

Better, over time, as your agents learn, forecasting should improve, resource allocation should sharpen and cross-market consistency strengthen.

In such a scenario, AI stops being a marketing experiment and becomes part of the revenue architecture.

The leadership question

For all of the reasons mentioned above, the question you should ask yourself can’t be “are we using AI?”, but rather “is our commercial operating model designed to learn?”.

Because complex B2B buying is collective and digital, intelligence must support how decisions actually form.

An AI-enabled commercial model continuously reads behavioural patterns, refines prioritisation, and aligns teams around shared commercial signals.

That requires structural design, turning AI into something far more strategic than a productivity tool.