AI adoption is widespread, but few marketing teams report transformational results
Nearly all marketing teams surveyed say they use artificial intelligence, but comparatively few describe its results as transformational. HubSpot’s 2026 State of AI in Marketing research found that 98% of respondents use AI in some way, while 10% said their teams achieve transformational results. The findings point to a challenge for marketers: moving beyond experimentation to repeatable workflows with measurable business outcomes.
The research surveyed 1,754 marketers across North America, Europe, Asia, Australia and Latin America. Respondents rated their own teams’ maturity, so the results reflect self-assessments rather than an independent audit of performance. The figures should be treated as directional benchmarks, not proof that AI use caused better results.
HubSpot defines AI maturity as a team’s ability to turn AI use into measurable business outcomes consistently. In marketing, that means more than counting tools or pilots. It includes documented workflows, access to appropriate customer data, human oversight and measurement against goals such as marketing-qualified leads, customer acquisition cost and revenue.
Four stages from experimentation to redesign
The research groups marketing teams into four stages: Early, Emerging, Established and Transformational. At the Early stage, AI use tends to be individual and informal. Emerging teams have a small number of owned use cases, but these may operate outside core workflows and be measured mainly by time saved or output volume.
Established teams have repeatable AI-supported workflows connected to customer data and tracked against funnel metrics. At the Transformational stage, teams redesign workflows around AI, with defined human review and governance. HubSpot’s framework emphasizes that teams can sit at different stages for different workflows, making it more useful to assess each use case than to assign one maturity label to an entire organization.
Across the survey, 42% of marketers said their AI use was advanced, meaning Established or Transformational. The comparable figures were 27% for sales and 23% for service. Among marketers surveyed, 96% of respondents identifying as Transformational said they met or exceeded their marketing-qualified lead targets, and they were roughly 3.5 times more likely to report outperforming customer acquisition cost targets. These are self-reported associations, not evidence that maturity alone produced the outcomes.
Assess workflows, data and governance
HubSpot recommends evaluating AI use cases across six areas: use cases, data and CRM context, workflow integration, governance and oversight, measurement, and strategy and ownership. Its suggested assessment scores each area from one to four, with evidence supporting higher scores. Examples include an inventory of owned use cases, documented access rules, process maps that identify human review, written policies and comparisons against a baseline or holdout group.
The framework also advises teams to focus on their weakest dimension. A marketing organization with many production use cases but limited governance, for example, may have a risk or oversight gap rather than a mature program. The assessment is a practical framework presented in the guide, not an official HubSpot metric.
Data readiness is another central consideration. AI workflows need accurate records, clear definitions, consent and permissions, appropriate access controls and an accountable data-quality owner. HubSpot’s cross-functional research found that 74% of companies using AI and reporting excellent data also exceeded revenue targets. The guide cautions that this is a correlation, not evidence of cause and effect.
Email nurture offers a measurable testing ground
For email marketers, the framework suggests connecting personalization and nurture workflows to customer relationship management data, then measuring performance against a defined objective. A team might establish a baseline for its existing nurture process, introduce an AI-assisted workflow to a portion of the audience, and compare results with a holdout that continues to receive the existing experience. The guide identifies a randomized holdout as the cleanest of its suggested options for email and nurture testing.
Measurement should match the workflow’s intended outcome. Teams can track an existing lead-nurturing metric alongside operational measures such as review time, error rate and cost per output. HubSpot notes that a change in one metric does not by itself establish that an AI workflow caused a revenue shift; attribution is complicated by other marketing touches and external factors.
The research also describes a four-part operating approach called Loop Marketing: Express, Tailor, Amplify and Evolve. It is presented as a way to move from pilots toward repeatable work, with people retaining judgment and oversight. The accompanying guidance recommends starting with one workflow, a named owner and a clear metric, then reassessing progress periodically.
For marketing leaders, the findings shift the practical question from how many AI tools a team has adopted to whether those tools are embedded in accountable processes. In email and across the wider funnel, that means connecting data, workflow ownership, review and testing—then treating reported gains as provisional until measurement supports them.
