Most recent surveys put marketer AI use somewhere between two thirds and nearly nine in ten, and the exact number depends entirely on what the survey counts as "use." HubSpot's 2025 AI Trends for Marketers report found 66% of marketers globally use AI in their roles, while SurveyMonkey's 2025 research put current-role or day-to-day use at 88%. Neither number is wrong. They are measuring different things, and if you've ever tried to answer how many marketers use AI in a single sentence, you already know it is not that simple.
This piece pulls together the surveys that actually publish their numbers, tells you what each one is really measuring, and reconciles the parts that look like they contradict each other. Read enough AI adoption statistics from different publishers and the disagreement starts to look deliberate, but it usually just comes down to a different question with the same word attached to it. No productivity claims, no ROI math, just what the data on AI adoption marketing teams shows about who is using AI, for what, and how that changes with team size.
What percentage of marketing teams use AI?
There is no single figure here, because there is no single question being asked. A handful of large surveys have tried to answer it, and they land in different places because they define "use" differently.
- 66% of marketers globally say they use AI in their roles, per HubSpot's AI Trends for Marketers report, updated June 11, 2025, based on more than 1,000 marketing and advertising professionals worldwide.
- 88% of marketers say they rely on AI in their current job or day-to-day work, according to SurveyMonkey's 2025 marketing AI research. SurveyMonkey does not publish the respondent count or fieldwork dates for this figure.
- 87% of content marketers say they use AI to create or help create content, per Ahrefs' June 2025 survey of 879 marketers (769 of whom reported using AI for content).
- 75% of marketers have adopted AI, according to Salesforce's 2026 State of Marketing research, based on a survey of 4,450 marketing decision-makers. Treat this as the broad current benchmark, but only with that attribution attached.
- 63% of marketing professionals use AI at least once a week, per ZoomInfo's 2025 State of AI in Sales and Marketing report, drawn from 1,002 US sales and marketing professionals across B2B and B2C companies.
Read those five numbers as five different questions, not five attempts to measure the same thing. The 88% figure answers "have you used AI in your job." The 63% figure answers a stricter question: "do you use it at least weekly." A marketer can answer yes to the first and no to the second if they tried a tool once and never went back. The 87% figure is not a marketing-wide number at all. It is specific to content marketing, and using it to describe all of marketing overstates what Ahrefs actually measured.
If you need one number to answer "what percentage of marketing teams are using AI right now," 75% from Salesforce's 2026 research is the closest thing to a broad, current benchmark, as long as you keep its source attached. For individual marketer usage specifically, 88% from SurveyMonkey is the more accurate figure to reach for.
What are marketers actually using AI for?
Across every survey in this space, content work is where AI adoption marketing teams report is heaviest and most consistent. Ahrefs found that among marketers using AI, 76% use it to brainstorm topic ideas, 73% use it to build content outlines, and 67% use it to improve existing content. Blog posts were the most common AI-assisted output, reported by 87% of respondents. AI-generated images sat at the bottom of Ahrefs' list at 19%, alongside translation and localization at 22% and keyword clustering at 23%.
HubSpot's numbers tell a similar story from a different angle. Half of marketers surveyed use AI for email and newsletter copy, and half use it for text-based social posts. Forty-seven percent use it for blog posts and other long-form content, 48% use it to generate marketing images, and 48% use generative AI for research. HubSpot's 2026 State of Marketing report puts content creation use even higher, at 80%, though that report does not disclose its sample size.
ZoomInfo's report on the topic found the same lean: marketing teams in its survey were most likely to reach for AI on content-creation tasks over any other use case.
SurveyMonkey's breakdown adds a few more use cases into the mix: 51% use AI to optimize existing content (including email and SEO), 45% brainstorm with it, 43% use it to automate repetitive tasks, and 41% use it for data analysis. Seventy-three percent say AI plays some role in personalizing the customer experience.
None of these percentages should be added together. Respondents in every one of these surveys could select multiple uses, so a marketer counted in the 51% who optimize content with AI might also be counted in the 45% who brainstorm with it. The honest summary is that AI use is functionally concentrated in content and the tasks that sit right next to it, not evenly spread across every marketing function. Image generation, translation, and structured tasks like keyword clustering lag well behind writing-adjacent work in every survey that breaks out use cases.
Does team size change adoption?
Yes, but not in the direction you might assume from "bigger companies have more resources." The pattern actually runs in three different directions depending on what you measure.
Larger organizations show stronger formal adoption. SurveyMonkey found that 57% of marketing teams at enterprises with more than 1,000 employees were willing to use AI, compared with 40% at companies under 1,000 employees. McKinsey's 2025 State of AI survey, fielded July 16 to July 31, 2024 across 1,491 respondents in 101 countries, found that 78% of organizations use AI in at least one business function, and marketing and sales was one of the functions where generative AI use was most common, at 42%. Companies over $500 million in annual revenue were more than twice as likely as smaller peers to have dedicated AI teams and a defined adoption roadmap.
Small teams are not simply behind, though. Ahrefs found AI use was actually most prevalent among companies with 10 to 49 employees and among enterprises with 1,000 or more, with the lowest prevalence at micro-companies of 1 to 9 people. Even among marketing teams larger than 100 people, 89% of respondents in that survey said they used AI. Social Media Examiner's small-business-weighted 2026 sample, where 60% of respondents came from businesses with 10 or fewer people, found 73% used AI daily and 90% used it at least weekly, though the piece does not disclose a sample size to check that against.
So the honest read is: bigger companies formalize AI adoption faster, with dedicated roles and roadmaps, while individuals at small teams often adopt tools directly and personally, without waiting for a program to exist. Neither group is simply "ahead." They are ahead on different things.
Why the numbers do not line up
Once you line every figure in this piece up side by side, the disagreement stops looking like sloppy research and starts looking like six different questions wearing the same word.
Different thresholds for what counts as use. "Ever tried an AI tool," "uses AI day to day," "uses AI at least weekly," and "company has fully implemented AI" are four separate things, and each produces a lower number than the one before it. SurveyMonkey's own data shows this inside a single report: 56% say their company is actively implementing AI, but figures the same article attributes to Salesforce put full implementation at 32% and active experimentation at 43%. An employee can use an AI tool the company never approved. A company can be experimenting with AI without deploying it to the whole marketing team.
Different populations. HubSpot surveyed 1,000-plus marketing and advertising professionals worldwide. Ahrefs surveyed 879 marketers, weighted toward content marketing. ZoomInfo surveyed 1,002 US go-to-market professionals across sales and marketing. McKinsey weighted its 1,491 responses by each country's share of global GDP. Social Media Examiner's sample skews heavily toward businesses with 10 or fewer employees. None of these are the same group of people answering the same question, so treating their outputs as one trend line is the mistake, not the numbers themselves.
Missing methodology. Several of the reports cited here, including SurveyMonkey's and Ahrefs', do not publish exact fieldwork dates or full sampling details on the page. That is not a reason to throw the figure out, but it is a reason to say "not disclosed" rather than implying a nationally representative sample that was never claimed.
Vendor-published research. HubSpot, SurveyMonkey, Ahrefs, ZoomInfo, and Salesforce are all commercial publishers with their own products to sell. Their numbers are still useful data points, but naming the organization behind every figure, the way this piece has, is what keeps a data-brief honest instead of turning it into an unattributed industry census.
What this means for your own team
If you are trying to figure out where your own marketing team sits, the honest answer is that the raw adoption percentage matters less than which of these questions you are actually asking. If the question is "has anyone on my team touched an AI tool," you are probably already past 66%, since that bar is low. If the question is "has my team moved past ad hoc use into something we actually manage," you are in the smaller group: the 32% SurveyMonkey cites as fully implemented, not the 88% who have simply tried a tool.
That gap, between individual marketers picking up AI tools on their own and a team running AI use through any kind of shared process, is where most of the friction shows up day to day. It is also where a platform that keeps brand voice, product facts, and past output in one place earns its keep, since it is the difference between five people improvising with five different tools and a team working from the same context every time. DeepSmith builds that kind of shared context once and applies it across research, drafting, linking, and metadata, which is one way to move from "someone here uses AI" toward a team that has actually implemented it. You can try it with a 7-day free trial before deciding whether it fits how your team works.
Wherever your team sits on that spectrum right now, the surveys above suggest you have company. Most marketing teams are somewhere between experimenting and formalizing, and the data does not support treating either end as unusual.
