In 2026, marketing teams aren’t short on AI tools.
What we’re still short on is AI infrastructure: the workflows, data foundations, and operating discipline that make AI usable at scale, without turning everything into generic noise.
Because when output gets easier, the gap between teams with strong foundations and teams without them doesn’t disappear. It widens. As Alex Bacon, AI Advisor and Founder of BrightKeel.AI, put it:
“What good looks like has not fundamentally changed.”
That single line captures the current state of marketing AI. High volumes of low-quality content still fall short. Vague positioning still underperforms. Weak data still leads to weak decisions.
The fundamentals still decide what works: strategy, audience understanding, positioning, brand, and measurement. But the pace has changed. And when speed increases, the consequences of getting the foundations wrong compound much faster.
This is the moment marketing becomes a systems discipline. Not because creativity is less important. But because creativity without infrastructure is now a bottleneck, and infrastructure without creativity is now dangerously easy to scale.
Nicole Leffer, CMO AI Advisor at A. Catalyst LLC sees a massive range in how teams are adopting AI, from “barely touched it” to “handing off entire workflows”. She said:
“There is an insanely wide range of where companies are.”
At one end: teams dabbling with free tools, personal accounts, and ad-hoc prompting - usually without training, standards, or shared workflows.
At the other end, organisations building autonomous, agentic workflows, sometimes using tools like Claude Code, and starting to talk about headcount reductions in a very real way.
“I am starting to see companies build out autonomous AI workflows where they are starting to genuinely hand off a lot of work - and not from the pie-in-the-sky perspective - but from the ‘we really don’t need this many people on our marketing team’ perspective.”
Nicole is clear that she wouldn’t recommend living on either extreme.
The sweet spot is the middle ground: teams that choose a foundational toolset, invest in training, and build repeatable workflows with humans in the loop.
The reason is simple. AI on its own doesn’t create an advantage. It accelerates whatever already exists. Without human judgment, strategic context, and clear standards, it simply accelerates average thinking.
But without AI, humans become the bottleneck in a market that now moves at machine speed. As Nicole puts it:
“The ones who are really, really excelling with AI genuinely see it as a partner, not there to do their work for them.”
The future of marketing isn’t man or machine. It’s man with machine - systems powered by AI, directed by humans, and grounded in data that can actually be trusted.
If you zoom out, both Alex and Nicole are saying the same thing in different words:
AI doesn’t replace marketing. It reveals marketing. It reveals whether you have:
Alex sees a common trap: teams try to bend their marketing around whatever the newest AI tool can do.
“One of the core principles is to look at how AI fits with you, your business and your marketing strategy, rather than trying to fit it to AI just because there’s a new tool that does something.”
In other words, the strategy still comes first. The tool comes second.
This is where many teams get it backwards. They adopt a new model, or a new workflow builder, or a new “agentic” tool, and then try to retrofit their marketing around its capabilities. The result isn’t a transformation. It’s distortion.
AI readiness isn’t a feature you switch on. It’s not a line item in a tech stack. It’s a team-wide competency. It’s the ability to:
Nicole reinforces this when she talks about what’s actually missing in most teams:
“A lot of people recognise that prompting skills are important. What I think a lot of people miss is really understanding what the tools are capable of.”
Without that understanding, teams either:
Both are symptoms of the same issue: shallow adoption. The deeper shift is this:
Marketing teams need to stop asking, “What can this tool do?”
And start asking, “Where does this tool strengthen the way we already create value?”
It’s about having the right foundations and a team mature enough to use AI without outsourcing their judgment to it.
One of the clearest signals that a marketing org is maturing with AI is the shift from “asking for outputs” to “designing systems.”
Alex used a simple example: the classic prompt, “write me a press release.”
“They’ve given AI a task, write a press release and it spits it out. And they go, ‘oh, this is great’, and then you read it and you kind of go, ‘that’s a bit rubbish really, isn’t it?’”
The problem isn’t that AI can’t help. It’s that the request is too big, too vague, and missing the thinking that makes marketing effective. Alex’s solution is the blueprint for AI-enabled marketing.
“They need to be able to break up tasks into specific workflow actions.”
Instead of “write the press release,” you split the job into steps AI can support:
That’s when AI stops being a shortcut and starts being a multiplier. And it’s exactly the same model Nicole recommends for operationalising AI across a marketing org.
“You need to have an understanding of your workflows. Like, what even happens?”
Nicole’s point is brutally accurate: most teams don’t know their real process because they’ve internalised it as muscle memory.
They think the workflow is “write a blog post, publish it.”
But the actual workflow is a chain: topic selection → research → SME input → transcript extraction → outline → draft → edit → distribution → measurement → iteration.
“So many marketing teams have no clue because a lot of it is just… they just do it.”
And that creates a ceiling on AI adoption, because:
“You can’t operationalise and scale something you don’t even have a process for.”
Again, this isn’t about getting AI to do everything for you.
It’s about understanding the mechanics of what you’re asking it to do, so you can get each phase of the process right and then streamline intelligently.
AI-ready marketing isn’t lazy marketing. In fact, it’s often the opposite.
If you’re going to do this properly, it will take more time up front to design the workflow than it would have taken to just complete the task manually. You have to break the work down. Define the inputs. Clarify what “good” looks like. Build the prompt. Test it. Refine it. Pressure-test the output.
That’s not shortcut behaviour.
But once the workflow is nailed, once the thinking is embedded and the guardrails are in place, that’s when AI starts to shine.
The teams willing to put in the upfront work to build the right foundations before chasing speed are the ones who will compound the fastest over time.
Most marketers use AI like a conversation. They open ChatGPT or Claude, type a request, tweak it, refine it, adjust it, and keep chatting back and forth until they get something usable.
Nicole sees this constantly:
“The vast, vast majority of marketers are using it in a way that they’re just chatting back and forth until they get the output that they want. But you can’t do that and scale it.”
And that’s the key. Chatting might work for a one-off task. It doesn’t build capability. The future state isn’t “more chat.” It’s reusable prompt systems and repeatable workflows.
Nicole’s advice is surprisingly simple:
“When you would normally say like, ‘Hey, change it this way’ use the edit prompt button and just edit your original prompt so it has that as part of the directions.”
Instead of refining the output in conversation, refine the instructions.
She suggests two practical techniques:
This is the hidden skill behind AI leverage in marketing: not prompting once, but building prompts that behave like SOPs.
And just as there is a responsibility for people to learn to prompt well to get AI to work effectively for you, there is also a responsibility to check and verify the quality of the output.
When teams say “AI isn’t working,” the root cause is often behavioural, not technical.
“People don’t take responsibility for what the AI created, they don’t review the outputs, and then it becomes whoever is your approver’s problem. In some cases, they don’t know what good looks like, and so they don’t know that what the AI gave them is bad.”
If your team can’t articulate what “good” looks like, in messaging, structure, insight, and tone, AI will amplify the ambiguity. It makes it easier to ship work without having to think.
The infrastructure conversation always comes back to data. Because AI isn’t magic. It’s pattern recognition. And patterns depend on inputs. Alex gives the simplest version of the truth:
“Rubbish in, rubbish out.”
But he also points to what’s changed: we can now connect and cross-analyse data in ways most teams never operationalised before. Historically, targeting was linear:
“I want to go after legal firms’ marketing managers, so I’ll pull a list.”
Now, teams can paint a richer picture: firmographics, technographics, funding, leadership structure, hiring, acquisition intent, and then tier outreach based on likelihood to convert.
“We’d pull together those different data sources and just be able to paint a little bit more of a picture of who is most likely to actually convert and engage. So you can then tier your outreach.”
This is where AI infrastructure becomes a competitive advantage in marketing: not because you have “more data,” but because you have connected context. And crucially, Alex reframes integration as the middle step, not the final step:
“That’s almost the easy bit in some ways, you’d work with a solution provider or someone in the tech team. The marketing part is either side of that: what data is valuable, and what are we going to do with that?”
This is the real question for marketing AI: What data do we trust enough to let AI influence our decisions? And where should those insights show up - SDR workflows, email automation, campaign prioritisation, ABM plays?
Both Alex and Nicole circle the same leadership challenge: how do you train teams to use AI without becoming lazy or over-reliant?
Alex frames it as a leadership and culture issue:
“It’s about leadership and culture in reality.”
The key is accepting that building good systems takes longer at first:
“It’s encouraging and accepting people to take that route, to take the longer route to accept that it’s iterative.”
Alex also offers a powerful alternative to outsourcing thinking: use AI as a coach.
“You can outsource the thinking to AI, or you can use it as a coach, challenge me, challenge my assumptions, ask me questions.”
Nicole’s version of upskilling starts with standardisation:
“You want everybody on your team to learn and use the same tool because then it becomes a team effort.”
Then she emphasises vetted training, because “learn AI” is not a useful directive.
“There’s so much noise and nobody even knows where to look, if you just like the average person, it’s like, ‘learn AI.’ They don’t know what to do with that direction.”
Her most important point is what training should actually include:
“A lot of people recognise that prompting skills are important. What I think a lot of people miss is really understanding what the tools are capable of.”
Because when teams don’t understand capabilities, they tool-hop endlessly, mistaking “new tool” for “new capability”, when the tool they were using already could probably do what they needed.
AI will keep improving. Autonomous workflows will get easier. Content production will get faster. But none of that changes the underlying truth.
AI is a multiplier. And multipliers don’t create value on their own. They amplify whatever they’re given.
If your workflows are unclear, AI scales confusion. If your positioning is weak, AI scales vagueness. And if your data is incomplete, outdated, or misaligned to your ICP, AI makes the wrong decisions.
That’s why the real competitive edge isn’t just adopting AI. It’s building the infrastructure that makes AI usable, reliable, and commercially intelligent.
Because AI doesn’t decide:
The organisations pulling ahead are those feeding AI clean, compliant, connected data and embedding it into repeatable workflows across marketing, sales, and RevOps. They understand that AI-ready marketing starts with data-ready foundations.
Trusted contact data. Accurate firmographics and technographics. Timely buying signals. Connected context across systems.
The future of marketing won’t be won by whoever writes the most content, launches the most campaigns, or automates the most workflows. It will be won by the teams who combine:
Into a single operating model.