How to win the content game in 2026 

“Talent wins games, but teamwork and intelligence win championships.” 

— Michael Jordan 

I’ve always liked this Michael Jordan quote, and weirdly enough, I think it applies pretty well to the content game in 2026. 

If I do say so myself, there are a lot of talented people working in content teams, and there are now more tools than ever to help them produce things faster. But before we get ahead of ourselves, talent and tools only get you so far in 2026. If the foundations underneath them aren’t right, you don’t really scale - you just create more work, faster. 

That’s why I keep coming back to something I saw at Forrester’s B2B Summit this year. They showed a very simple 2x2 grid that asks two questions about any piece of work:  

Before deciding who should do the work, or where AI should come into it, you first ask: 

What kind of work is this? 

It sounds almost too simple, but I think it gets to the heart of one of the biggest problems we’re seeing with AI and content right now. 

The Truth is: AI hasn’t automatically made content faster 

 
More than ever, there’s enormous pressure on content teams to produce more, and the gap between those who create and those who order content is widening.  

Forrester’s State of B2B Content Survey found that 51% of content professionals say leaders expect more content, faster, because they believe AI makes that possible. Among leaders themselves, that rises to 57%, compared with 47% for everyone else. 

I find that gap interesting because it means the people setting the targets are more convinced that AI should make everything faster than the people actually trying to make it work. 

And from what we see with clients, that makes sense. 

We work hands-on with marketing and content operations across technology, cybersecurity, manufacturing, financial services and many other B2B sectors. When AI enters those environments, something slightly counterintuitive often happens: content doesn’t immediately get faster. Sometimes it actually gets slower. 

The issue isn’t that AI can’t produce a draft quickly. Of course it can. The issue is everything that happens around that draft.  

Those are all reasonable questions, especially in industries where getting something wrong can have very real consequences. The problem starts when those questions become part of the workflow for virtually everything. 

Suddenly legal is reviewing a nurture email, product is checking a routine LinkedIn post and an SME is being asked to approve a campaign adaptation that barely changed. AI may have created more drafts, but it has also created more things for people to review. 

The guardrails aren’t the problem 

I don’t think companies are wrong to be cautious here. A cybersecurity vendor can’t publish a detection claim just because an AI model produced it confidently, and a bank can’t let a model invent offer terms. Legal, compliance, product and subject matter experts all exist in these workflows for good reasons. 

The mistake is applying the strictest possible guardrails to everything. 

That’s what I like about the Forrester framework. It gives you a relatively simple way to separate the work based on its actual risk and complexity. Product claims, pricing or regulated content might be relatively straightforward to create but carry significant risk. Core positioning, executive thought leadership, and major campaign narratives carry both risk and complexity. Routine follow-ups, campaign adaptations, and social extensions sit somewhere very different. 

The work is different, so the process should be different too. That makes sense – but Forrester went one step further. 

Ownership should follow the work 

Forrester claimed that ownership follows the work, not the org chart. 

That line stuck with me because so many content processes still start from the opposite position. We ask which team normally owns an asset and then send the work through that team’s established process, regardless of what the work actually requires. 

Forrester flips that around. High-risk, high-complexity work stays close to experienced content professionals, domain experts and SMEs. Lower-risk execution can move closer to marketers and domain teams doing the day-to-day work, while complex production may make more sense with agencies or specialist partners. 

Instead of asking, “Which department normally does this?”, you ask, “Who is best equipped to do this particular piece of work?” 

It’s a small change in thinking, but at scale it can make a big difference. 

Set the defaults instead of debating every asset 

I’d take the framework one step further. 

You don’t want five people sitting in a meeting deciding which quadrant every new email belongs in. The better approach is to establish sensible defaults for recurring content types and build those decisions into your intake and production workflows. 

That may sound pretty complicated, but let me explain through a few examples. 

Once you’ve established those defaults, the classification can determine who owns the work, whether AI can draft it, how much review it needs and when legal, product or an SME becomes involved. 

Then you define the exceptions. 

If that normally low-risk webinar email suddenly contains a new product claim, customer data, pricing or regulated language, it moves into a stricter workflow. Maybe it includes product marketing and leadership to be involved.  

The point with exceptions is you don’t need to reconsider the entire system every time; you just need clear triggers for when the normal process no longer applies. 

The key is: Classify by default. Escalate by exception. 

That’s where I think this becomes genuinely useful as an operating model rather than just an interesting Forrester slide. 

But classification only solves half the problem 

Even if you know where AI can safely help, there’s still another pretty important question: 

What does the AI actually know? 

The Forrester framework helps with governance and ownership, but good execution still depends on context. This is where it connects with something I presented at our Zen Breakfast with Clay in London earlier this year: layer intelligence before execution.