A few weeks ago I was reading OpenAI launch article for workspace agents when a detail stood out. They described the state of most companies’ AI adoption as “scattered prompts and half-constructed workflows” accumulated over two years. It’s a reality check straight from the company that sparked the AI boom in the first place.
Despite the headlines, most organizations still haven’t fundamentally changed the way they work. AI can generate solid content in less than a minute, but faster tasks have not automatically created faster organizations.
Most marketing teams have spent the last two years doing exactly what they were incentivized to do. Each specialist has figured out how to make AI useful in their own workflow.
The content specialist writes newsletter excerpts in ChatGPT. The designer generates on-brand graphics in Firefly. The email marketer created a quality assurance workflow that saves hours every week. Managers use ChatGPT to verify the integrity of the copy before shipping.
This is real progress. But it is also unrelated progress.
Your customers are searching everywhere. Make sure your brand introduces himself.
The SEO toolkit you know, plus the AI visibility data you need.
Start free trial
Start with
AI hasn’t solved the workflow bottleneck
In my upcoming book, “Hyperadaptive: Rewiring the Enterprise to Become AI-Native,” I review a composite marketing team I call Meridian Digital. They are responsible for turning a Monday morning blog post into a Tuesday email newsletter. In the pre-AI world, this work took four days.
With each person using AI in their work, individual tasks became faster, but producing the newsletter still took four days. Transfers are always human. Expectations between people are still human. Approvals are always human. Everyone was 30% faster on their individual task, and the overall process remained the same.
This is what local optimization looks like on an organizational scale. This model is not unique to marketing: any specialist team that has absorbed AI individually is likely to be somewhere close to this workflow.
When I share this example with marketers, most say something like, “That’s almost exactly where we are. »
How to link specializations together
The mistake I see organizations making is assuming that this bottleneck is a tool problem. This is not the case. Agents Workspace, marketing agent platform from Jasper, Copilot and Claude Skills. All of these are significant advancements at the platform level, and they all give you a place to put the good work.
But the job of connecting specializations is not a feature you can buy. This is a structural change, which is why so few of the 90% have moved to the 10% who are seeing real impact with AI.
When was the last time your team took an honest look at how much of your AI progress is within specializations rather than across them?
I mapped the journey organizations take from initial AI confusion to becoming AI native. It has six waypoints, and the most important part for this conversation is the middle one.
- AI Bifurcation: Power users are ahead of everyone else. Strong use cases are emerging in pockets. Progress stalls because the rest of the team is still figuring out what AI means for their role.
- Localized progress: Individual automations are starting to appear – one for content, one for design, one for reporting – but they are siled. There is no repeatable way to develop them within the team.
- Coordinated progress: A network of AI activation hubs is forming. Automations connect. The organization begins to behave like a native AI system rather than a collection of AI-enabled individuals.
Most marketing teams I work with fall somewhere between bifurcation and localized progress. They are not yet in coordinated progress because no one has been asked to take ownership of the work of connecting the specializations.
You can’t buy your way past this gap. You have to build it. It’s a role (I call it an AI Lead), a model (the Enablement Hub), and a cadence that transforms winning workflows into shared infrastructure, reviews them, removes what doesn’t work, and improves what works.
The Meridian Digital newsletter workflow is very different when coordinated progress is being made. In this scenario, the system detects the new blog post, writes three variations of the newsletter, generates on-brand graphics, performs quality checks, and presents options to the manager during their regular content review.
The delivery time goes from 4 days to 1 day. Cycle time goes from two hours to one hour. Jobs are changing. The content specialist becomes an orchestrator of the content system. The designer moves on to visual direction and guardrails. The email marketer starts creating the automations themselves.
Your next move
I leave you with this challenge. Look honestly at the map. Most of the marketers I work with discover that they are at an advanced stage of where they thought they were, and that discovery is what makes the next step obvious.
