The 36-point gap between AI leaders and everyone else

A recent study from ForgeX found that top-performing ABM teams are nearly three times more likely to have a documented AI roadmap than everyone else. 59% of top performers have one. 23% of the rest do.

The gap isn't in tools. The tools are the same. Every marketing team has access to the same AI features in the same platforms. Predictive scoring, automated segmentation, content generation, send-time optimization, intent signal processing. The technology is available to everyone.

The gap isn't in budget. AI features are increasingly included in platform licences. The cost of entry is falling, not rising.

The gap is in documentation. The top performers wrote down what they're trying to achieve with AI, which use cases they're prioritizing, how they'll measure success, and what governance applies. Everyone else activated features and hoped for the best.

That distinction, between deliberate deployment and hopeful activation, is the difference between AI that produces measurable results and AI that produces activity nobody can connect to outcomes.

Why documentation is the differentiator

This seems counterintuitive. In a technology-driven landscape, you'd expect the differentiator to be technology: better models, more sophisticated algorithms, more advanced platforms. Instead, the differentiator is a document. A plan written down on paper (or more likely in a shared drive nobody checks often enough).

But the document itself isn't the differentiator. What the document represents is.

A documented AI strategy means someone sat down and answered hard questions before activating anything. Which specific business problems are we using AI to solve? Not "improve marketing" but "reduce MQL-to-opportunity conversion time by 20%" or "increase scoring accuracy so that MQL-to-SQL acceptance rate exceeds 40%." Specific, measurable outcomes tied to specific AI capabilities.

A documented AI strategy means someone defined which AI features to activate and which to leave alone. Not everything available is worth using. Predictive scoring might be worth activating if the data is clean enough to support it. Automated content generation might not be worth activating if the brand voice requires human judgement. The document forces the team to evaluate each capability against their specific context rather than activating everything because it exists.

A documented AI strategy means someone defined how success will be measured. Not "is the AI running?" but "is the AI producing better outcomes than the non-AI approach it replaced?" Without a defined measurement framework, the team can't tell whether AI is helping, hurting, or making no difference at all. The 77% without a documented strategy are, by definition, unable to answer this question.

A documented AI strategy means someone assigned governance. Who owns each AI feature? Who monitors its output? What happens when it produces unexpected results? Who reviews it and how often? Without these answers documented, governance exists in theory but not in practice. The AI runs, nobody watches, and problems accumulate until they're too large to ignore.

What "undocumented AI" actually looks like

In the 77% of teams without a documented AI strategy, the pattern is remarkably consistent.

AI features get activated opportunistically. A platform update ships a new AI capability. Someone on the team turns it on because it looks useful. Nobody documents what it does, what data it uses, what it's supposed to achieve, or who owns it. The feature runs. Other work continues. Nobody checks back.

AI adoption is measured by activity, not impact. Leadership asks "are we using AI?" The team says "yes, we've activated predictive scoring, automated segmentation, and AI-assisted content." Leadership is satisfied because the answer was yes. Nobody asks the follow-up: "what measurable improvement has AI produced compared to our previous approach?" Because nobody set up the measurement to answer it.

AI creates confidence without evidence. The predictive model produces scores. The scores look precise. The team trusts them because the AI calculated them. But nobody compared the AI-scored leads against the manually-scored leads to see whether conversion rates actually improved. The confidence comes from the sophistication of the tool, not from evidence that the tool is working.

AI features conflict without anyone noticing. The MAP's AI-powered scoring adjusts lead scores based on one set of patterns. The ABM platform's AI prioritises accounts based on a different set of patterns. The CRM's AI recommends actions based on yet another perspective. Three AI systems, three different recommendations, no coordination between them. The sales rep sees contradictory signals and follows their gut because the AI isn't telling a coherent story.

Each of these patterns is preventable. Each one is prevented by the same thing: a document that answers the questions the team never asked before activating.

What a documented AI strategy actually contains

The document doesn't need to be long. It needs to be specific. Five sections, each one answering a question the team needs to align on.

Use cases. Which specific AI capabilities are we activating and why? For each one, what business problem does it address, what outcome do we expect, and how does it connect to a metric leadership cares about? If you can't connect an AI capability to a specific outcome, don't activate it yet.

Prioritisation. Which use cases are we doing first and which are we deferring? Not everything can be deployed simultaneously. The data might not support all use cases equally. The team's capacity to govern multiple AI features might be limited. Prioritise based on potential impact (which use case addresses the biggest gap) and readiness (which use case has the cleanest data and the clearest measurement framework).

Measurement. For each activated AI capability, how will we know it's working? Define the baseline (performance before AI), the target (performance we expect with AI), and the timeline (when we'll evaluate). Without a baseline, you can't measure improvement. Without a target, you can't define success. Without a timeline, you'll never get around to checking.

Governance. Who owns each AI feature? Who monitors output? What review cadence applies? What happens when something goes wrong? This section prevents the "activated and forgotten" pattern that plagues most AI deployments.

Dependencies. What needs to be true for each AI capability to work as intended? Is the data clean enough? Are the integrations reliable enough? Does the team have the skills to configure, monitor, and interpret the AI's output? Dependencies are the readiness check that prevents the team from deploying AI on a foundation that can't support it.

The document is cheap. The absence is expensive.

Writing a documented AI strategy takes a day. Maybe two if the team has a lot of AI features already active and needs to catalogue them first. The cost is trivial compared to every other marketing investment.

The cost of not having one is measured in AI features running without governance (compliance risk), AI making decisions on bad data (performance degradation), AI producing results nobody measures (wasted investment), and the team unable to answer when leadership asks "what is AI actually doing for us?" (credibility loss).

The 59% who documented their strategy aren't smarter. They're not using better tools. They simply answered the hard questions before deploying, which meant every AI feature they activated had a purpose, a measurement framework, and an owner. The 77% who didn't answer those questions activated the same features with no purpose, no measurement, and no owner. The results diverged accordingly.

The action is straightforward

If your team doesn't have a documented AI strategy, the next step isn't to build a complex framework or hire a consultant. It's to book a half-day session with the people who manage your marketing platforms and answer five questions together:

What AI features are currently active in our platforms? What data does each one consume? Who owns each one? How do we know if each one is working? What would we turn off tomorrow if we had to choose?

Write the answers down. That's your documented AI strategy. It's not polished. It's not comprehensive. But it's infinitely more than what 77% of teams have, and it puts you on the side of the gap where results actually follow investment.

The AI is available to everyone. The strategy is what separates the teams that get value from it and the teams that just run it.