Personalisation is one of the most cited AI use cases in B2B marketing, and one of the most frequently oversold. The technology has advanced genuinely. The infrastructure and data most organisations have has not.

McKinsey's 2023 personalisation research found that companies excelling at personalisation generated 40% more revenue from those activities than average performers. Gartner's B2B buyer research shows that buyers who receive personalised, relevant content at key decision points are more likely to complete a purchase and less likely to experience buyer's remorse. The case for personalisation isn't in dispute.

What is in dispute: whether AI makes personalisation easy, and whether most B2B organisations are positioned to benefit from it today. The answer to both is more nuanced than most vendor conversations suggest.

What the Research Shows About Personalisation's Effect on B2B Buyers

The evidence for personalisation's effectiveness is real, but it carries important qualifications.

McKinsey's 2023 "The Value of Getting Personalization Right" found that 71% of consumers expect companies to deliver personalised interactions, and 76% get frustrated when this doesn't happen. In B2B contexts, Gartner research shows that sales and marketing personalisation at key stages of the buying journey measurably improves both win rate and deal size, particularly when content maps to the buyer's role and stage in the decision process.

But the Ehrenberg-Bass Institute's body of work complicates the picture. Their research consistently shows that in categories where buyers are not currently in-market, distinctiveness and salience, being consistently present and recognisable, often outperforms relevance. For most B2B audiences, the majority of potential buyers are not in an active buying cycle at any given moment. Personalisation, which typically targets in-market behaviour, doesn't address them.

The practical implication: personalisation is a powerful conversion tool for in-market buyers. It is not a substitute for the brand-building activity that creates the pool of buyers who'll remember you when they enter the market. Both matter. AI tends to be discussed in the context of the first, while the second is often underfunded.

Where AI Personalisation Actually Works at Scale

The AI personalisation use cases that deliver genuine, scalable value in B2B marketing tend to cluster around four areas.

Email subject lines and content variation. AI can test, learn and personalise at a scale no human team can match. Segment-level subject line variation, content module selection based on engagement history, and dynamic send-time optimisation are all well-established use cases with strong evidence behind them. This is the highest-leverage, lowest-barrier entry point for most B2B teams.

Ad creative variation. Generating multiple versions of ad copy or headline variants for testing, and optimising ad delivery toward the audience segments where performance is strongest, is a mature AI capability. The marketer's job shifts from writing every variant to defining the strategy, brief, and boundaries within which the variation operates.

Website content modules. For organisations with sufficient traffic and a content library, AI-driven personalisation of homepage modules, resource recommendations, or case study selection based on visitor characteristics is achievable. The prerequisite is a content supply chain that can produce enough variation and a content management system that supports dynamic delivery.

Recommendations where purchase or engagement history exists. This is where AI personalisation is most proven, and where most B2B organisations have the least relevant data. If you have transaction history and engagement history per account, AI recommendation engines produce genuine value. If you don't, the recommendation engine has nothing meaningful to learn from.

The pattern across all four: AI personalisation works when it has clean, relevant, sufficient data to learn from. Where the data is absent or fragmented, the capability is limited regardless of the sophistication of the AI.

The Data and Infrastructure Requirements Most Organisations Don't Have Yet

Here is the part most AI personalisation conversations skip: the data and infrastructure requirements are significant, and most B2B organisations haven't met them.

Unified customer data. Personalisation at scale requires a single view of each account and contact across every touchpoint: website, email, CRM, ads, events, support. Most B2B organisations have this data distributed across multiple disconnected systems. Without unification, AI personalisation acts on partial signals and produces partial results at best.

Identity resolution. In B2B, the same person may interact with your brand across multiple devices, in multiple roles at the same organisation, across years. Stitching those interactions into a coherent identity is technically difficult and often expensive. Without it, personalisation treats the same buyer as multiple strangers.

Content supply chain. Personalisation requires content variation: multiple subject lines, multiple body modules, multiple landing page headlines, multiple ad creatives. Most marketing teams don't produce content at the volume required to make AI-driven selection meaningful. The bottleneck isn't the AI, it's the content production capacity behind it.

Forrester's 2024 B2B personalisation research found that fewer than a third of B2B organisations had the data infrastructure in place to support AI-driven personalisation effectively. The technology is available. The prerequisite investment often isn't in place yet.

How to Prioritise Personalisation Efforts Given Real Constraints

The mistake most B2B marketing teams make with AI personalisation is starting with ambition rather than inventory. They identify the most sophisticated use case and then discover they don't have the data or infrastructure to support it.

A more practical approach starts with two questions: where do we already have sufficient, clean, connected data, and where is the buying stage such that personalisation will meaningfully improve the decision or accelerate the conversion?

The answer is usually email, specifically email to existing contacts and accounts where you already have engagement history and CRM data. It's the highest-leverage, lowest-infrastructure personalisation use case most B2B organisations can execute now. Start there, build the capability, measure the impact, and expand from that foundation.

The natural expansion path is predictable: email to account-level website personalisation (if you have account identification capability), then to ad personalisation, then to multichannel orchestration. Each step requires incrementally more infrastructure and produces incrementally more value when that infrastructure is in place.

The complexity trap is the opposite approach: trying to build a unified data platform, implement identity resolution, and launch a multichannel AI personalisation programme simultaneously. That's a two-year infrastructure project that produces no marketing output for 18 months and often gets defunded before it delivers. Start where the data already is. Build out from there.

AI scales what's working. It doesn't replace the groundwork that makes personalisation possible in the first place.

Ready to Build Your AI Personalisation Strategy on Solid Ground?

The Marketing and AI track at FP Collectiv covers AI personalisation in depth: how to evaluate your organisation's readiness, where to start, and how to build out a programme that actually delivers. It's grounded in named research and designed for B2B marketers making real decisions about investment and priority.

No vendor recommendations. No hype. Just the evidence and the framework for making a better call.

Sources

McKinsey & Company, "The Value of Getting Personalization Right — or Wrong — Is Multiplying," 2023.
Gartner, B2B Buyer Research, various publications 2023–2024.
Ehrenberg-Bass Institute for Marketing Science, various publications on advertising effect and distinctiveness.
Forrester Research, "B2B Personalisation Survey," 2024.

Note: Aggregator-sourced figures should be traced to primary sources before publication.