The pattern is familiar across B2B organizations. Marketing generates a large volume of leads, sales concentrates on a much smaller group of accounts, and the two teams evaluate the pipeline through different dashboards and definitions. When revenue misses its target, each side can point to activity that appeared successful inside its own system.
Account-based marketing exists because that disconnect is expensive. Instead of beginning with the largest possible audience and filtering it down, ABM begins with a defined set of valuable accounts and coordinates marketing, sales, and customer-success activity around them.
The discipline is no longer a niche experiment. The 2025 Account-based Marketing Benchmark Survey reported that 71% of respondents were using an ABM strategy, while nearly 65% said their initiatives were meeting or exceeding organizational expectations.
The same survey also exposed the weakness beneath that adoption. Forty-five percent of respondents said they understood their ideal customer profile but still needed better account coverage. Only 33% said sales and marketing were aligned on that profile and had strong coverage of their target accounts.
That distinction is the core of ABM’s 2026 reckoning. A platform can score, enrich, rank, and personalize accounts, but it cannot decide which customers fit the company’s economics or force two departments to agree on what a valuable opportunity looks like.
The shift to an account-based model is partly a measurement shift. Reach, lead volume, and cost per lead become less important than engagement, pipeline movement, and closed revenue within the named-account group. That change can reveal how much activity was reaching people and companies that sales never considered credible buyers.
Awareness also takes on a different meaning under ABM. A complex B2B purchase may involve financial, operational, technical, and executive stakeholders, each evaluating a different risk. Reaching the company name is therefore not enough. The program has to reach the relevant people with messages connected to their particular role in the decision.
This is why personalization depends on selection. Tailored advertisements, emails, and landing pages can be useful when they are built for accounts that fit an agreed customer profile. If the list is poorly defined, personalization simply makes irrelevant outreach look more polished.
The profitability argument requires the same discipline. ABM concentrates more time and spending on fewer accounts, so its performance should be judged at the account and pipeline level rather than by raw lead totals. A sound list does not guarantee a return, but it is a necessary condition: no increase in production speed can make the wrong accounts more likely to buy.
The 2025 benchmark data illustrates the gap between AI’s promise and its realized value. Thirty-seven percent of respondents said they were using AI tools to support ABM, and 45% credited AI with improving personalization at scale. Yet nearly 70% rated AI as having little or no effect on campaign outcomes, while only 3% considered it a significant difference-maker.
The reported obstacles were operational as much as technical. Forty-five percent cited integration with the existing technology stack as a major difficulty, and 43% cited a lack of specialized internal skills. Across ABM more broadly, proving return on investment was the leading challenge at 47%, followed by sales and marketing alignment at 43%.
Those findings complicate the vendor narrative. AI can produce more variations, process more signals, and accelerate campaign execution. It cannot repair a disputed customer profile, an account list that sales will not use, or a measurement model that rewards marketing for volume while holding sales responsible for revenue.
The infrastructure supporting these capabilities is expensive. In its second-quarter 2026 results, Tencent reported capital expenditure of 52.8 billion yuan, up 176% from the same quarter in 2025 and approximately 65% from the preceding quarter. The company said it had substantially increased its procurement of computing capacity for AI models, applications, and services.
Tencent separately reported that marketing-services revenue rose 22% year over year to 43.6 billion yuan. It attributed that growth partly to enhancements in its AI-driven advertising recommendation model, upgrades to its automated campaign-management system, and tighter integration across the Weixin ecosystem.
Those figures demonstrate both sides of the AI economy: providers are investing heavily in computing infrastructure while using AI to improve commercial advertising systems. They do not, however, prove that every advertising platform will recover those costs through higher prices or that every buyer of AI-enabled marketing software will receive an equivalent return.
For an ABM team, the practical test is narrower. An email-generation tool may increase production speed, and a landing-page system may make account-level variation easier. An intent product may surface useful buying signals. Each tool still has to be evaluated against the agreed account list and against movement in qualified pipeline, rather than against the volume of content or alerts it produces.
An example from outside marketing helps clarify the role of implementation. A Microsoft case study about a Lima education pilot described nearly 500 primary-school teachers across more than 200 public schools receiving training in Copilot Chat. Marco Antonio Pedraza, a sixth-grade teacher featured in the article, learned to create personalized classroom activities by giving the system specific information and clearer prompts.
The Microsoft article documented individual teachers’ experiences, but it did not establish improved student outcomes. It said an impact assessment was expected later. That distinction matters because an encouraging case study is not the same as measured evidence.
Later reporting from the World Bank’s work in Peru offered a firmer operational result. A randomized trial across 390 schools involved 1,270 teachers and found that brief, practice-oriented training increased active AI use while helping participants write stronger prompts and evaluate outputs more critically.
The lesson for ABM is not that classrooms and sales organizations are equivalent. It is that access to a tool is different from the ability to use it well. Defined tasks, relevant inputs, training, and an evaluation framework determine whether faster output becomes useful output.
ABM therefore works best when a company can answer three questions clearly. Which accounts matter, and why? What does the buying group inside each account need to understand? How will marketing and sales recognize progress using the same evidence?
Companies that can answer those questions give AI a useful operating boundary. Companies that cannot may simply produce faster and more expensive versions of their existing confusion.
The regulatory backdrop reinforces the value of knowing who is being targeted and for what purpose. As state-level privacy obligations expand, marketers must navigate a growing collection of consent, profiling, disclosure, and opt-out requirements.
That legal expansion does not automatically make ABM cheaper than broad outreach. It does make poorly governed data collection harder to defend. A defined account strategy can give an organization a clearer reason for collecting data, selecting audiences, and limiting activity to the people relevant to a legitimate sales effort.
The harder question is whether a company has the organizational maturity to run ABM at all. It requires marketing and sales to share a definition of a valuable account and a common view of the pipeline. It also requires tailored content, realistic account coverage, disciplined data practices, and leadership willing to accept smaller activity totals in exchange for greater commercial relevance.
AI can accelerate research, personalization, and campaign production once those conditions exist. When they do not, its speed becomes a liability because weak assumptions can be repeated across more accounts and more channels before anyone tests whether they were right.
The companies positioned to earn stronger returns from ABM in 2026 will not necessarily be those with the longest tool lists. They will be the ones that decided which accounts mattered, secured agreement from the teams responsible for them, and gave every new AI capability a measurable job inside that strategy.