Closing the gap between AI expectations and reality in marketing
99% of marketers are using AI. 94% trust its outputs to some degree. And 65% say it makes them more productive.
One of the most interesting findings from our upcoming Secret Lives of Marketers report, based on research from 1,500 marketers across the UK, USA and Australia, is that AI adoption is almost universal across the marketing industry, but the relationship between marketers and AI is complex at best.
Only 36% of marketers say AI actually makes their job easier, and exactly the same proportion (36%) say they enjoy using it.
Marketers are using AI. But are we using it well?
The AI conversation has moved on; we’re no longer wondering if marketers should use AI; we’re now talking about adoption and impact and whether AI is actually delivering what marketers need. 65% of marketers say that AI speeds up their productivity, but 34% have a more mixed experience, saying it speeds up certain tasks and slows down others.
When asked how AI could slow things down, our survey pool pointed to:
- Learning how to use AI tools effectively (41%)
- Writing and refining prompts (38%)
- Dealing with inaccurate or unreliable outputs (36%)
- Fact-checking and verification (35%)
- Compliance and approval processes (33%)
- Editing and refining outputs (30%)
We know that AI can help marketers produce more, analyze more, and move faster. But it seems that it can simultaneously create a whole new layer of work around those outputs.
Marketers aren’t alone in experiencing this tension. BCG’s 2026 global AI at Work research found that 47% of workers now spend more time managing and directing AI than doing the work itself, with 41% of regular AI users saying it has increased their cognitive load.
A key takeaway here for marketing leaders is that BCG also found that a clear AI strategy increased AI’s positive impact by 25 percentage points, compared with just five percentage points from providing better tools.
Buying more AI tools isn’t a guarantee for scaling success within your team unless you have a clear plan on how to fit those tools into the way people actually work.
Download our free AI-friendly tone of voice template
To trust, or not to trust? That’s the AI conundrum
94% of marketers trust AI to some degree. That’s a huge figure, but it’s not the whole picture.
While 47% of marketers say they fully trust AI outputs, 48% only trust them after fact-checking, and 1% don’t trust AI at all.
The complexity around trusting AI is increased by 36% of marketers saying AI can produce inaccurate outputs. This isn’t particularly surprising given that even LLMs themselves include a warning that states mistakes can be made and recommend users check the outputs. In fact, one of the most famous AI hallucinations in existence never fails to raise a smile, but is also a great reminder of how easily things can be misinterpreted:
The things marketers spend most time fact-checking are, of course, the areas where accuracy is essential. A quarter (25%) of marketers spend the most time fact-checking AI-generated statistics and research, while 20% say the same about reports that use their own data.
This is not an argument against AI, just further evidence of the importance of using it well.
Has AI made marketing easier?
One of the most complicated questions in the world right now is about how/if AI is enhancing our lives. And that is no different in the marketing industry. Marketers are being asked to keep pace with changing customer behavior, privacy and compliance, new technology, data, content, attribution, and constantly evolving marketing channels.
This creates an expectation to understand creativity and psychology while remaining analytical and commercially minded. Rather than removing this added workload, AI has the potential to become another thing marketers have to manage, unless it is integrated seamlessly into an existing tech stack.
That’s why the next stage of marketing’s AI story needs to be less about having AI (we all have it) and more focused on making it genuinely useful.
Generic AI is not the solution
45% of marketers use AI for content creation, but 24% say AI-generated content is too generic.
And most marketers can easily recognize this problem. Creating ten variations of a campaign message in seconds sounds super efficient. But if those variations don’t sound anything like your brand, or speak directly to your customers’ needs – they aren’t adding value to your marketing strategy.
AI earns its value when it can work with the information marketers already have: brand voice, customer data, campaign strategy, previous activity and the workflows their teams actually use.
Dotdigital’s Content Agent and Segment Agent have been designed to do exactly this. They can be trained on your brand’s voice and guidelines, as well as the specific audience needs for each campaign you create.
The question marketers should be asking of any AI tool is quite simply: will this make my work better or just faster?
The marketing goal isn’t more AI. It’s better use of AI
With 99% of marketers already using AI, increasing AI adoption is no longer the goal.
The focus should now be on making AI easier to use, connected to the right data, and trained on the things that make your brand unique.
Establishing clear guidance around when AI should be used, when fact-checking becomes essential, and where human judgment cannot be replaced removes the AI uncertainty without slowing down experimentation and AI integration.
The marketers who will succeed the most with AI won’t be the ones who are using it the most, but the ones who know when, where, and how to use it for the biggest impact, without losing the creativity and expertise that makes marketing human.
Check out Dotdigital’s 2025 Secret Lives of Marketing teams report for more AI stats, and stay tuned for the 2026 version
