Opinion: faster AI can highlight and worsen marketing workflow bottlenecks

Marketing leaders expected Claude’s rollout to create a giant productivity leap, but some six months later, little has changed. As a consultant, I have to explain that tools alone don’t create efficiency — people do.

That’s because AI isn’t a plug-and-play efficiency layer; instead, it’s a mirror. Adding AI to your workflow will certainly amp up the processes that already work well. However, any broken processes that might have been overlooked are likely to spread like wildfire. If you’re working with an operating model that’s 10 or 20 years old, you’re not maximizing the ROI on your technology investment. Let’s take a look at what that means and how to fix it.

AI improved output times, so why did delivery lag?

Your marketing team has probably always had bottlenecks in its workflow, but now that AI moves work along so much faster than before, those bottlenecks stand out like they’re wearing an ’80s neon tracksuit and doing the Hammer dance.

For example, I know an insurance content team that recently purchased an AI tool, expecting it to improve content delivery fivefold with half the staff. It kind of worked. They could easily produce much more content, but they couldn’t deliver it any faster than before.

The team produced a boatload of content, but then it got stuck in approvals. The approval process didn’t change. With Claudette (a real-life human, not a feminized version of Claude) expected to approve five times as much content, she became an even bigger bottleneck than before. She was so overwhelmed by the volume of approvals that delivery slowed down.

The silver lining in this story is that my team facilitated a conversation between Claudette and her team to discuss the bottleneck and make it clear to everyone that something needed to change.

Skip ahead a few months, and the team built a first line of approval with AI, reducing Claudette’s review to a final spot check of cleaned-up copy.

The world’s most powerful SEO platform, purpose-built for Enterprise.

Three common workflow challenges AI exposes

Now that AI tools have been widely used by marketers for a couple of years, I’m seeing three common workflow challenges with clients: unclear approval chains, no single source of truth for the brand, and undefined data and measurement ownership. Let’s dive into how you can fix them.

Unclear approval chains

AI drafts work quickly, but with no single defined approver — or, conversely, five stakeholders each holding veto power — speed gains evaporate into email threads.

For example, we worked with a homebuilding company developing a GTM campaign for a new-home community. The team’s most productive employee (let’s call him Chat) cranked out flyers, signage, and social media assets before I could seemingly even count to 10.

That sounds amazing, but it caused a lot of confusion around approvals. All the work was ready for review the same day, but there was no clear process for who owned the approvals. The leadership team spent three weeks going back and forth about that. So much for the speed of AI.

However, that delay ended up being a good thing because it exposed an issue that went unnoticed when processes were moving more slowly. The leadership team updated their processes, knows who approves what, and implemented a 24-hour turnaround policy.

Disagreements around brand standards

Here’s another situation I’ve encountered. The brand team at a healthcare company designed collateral for the grand opening of its new hospital. They were elated because, with the power of AI, what normally took several weeks to create was done in just a few days. What a win, they thought. But when it was time to approve the work, the team began arguing over the use of colors and fonts. The nasty truth was laid bare — the company had never documented any brand standards.

After a few collaborative workshops, the team agreed on brand standards. Now, reviews happen with fewer disagreements, improving delivery times.

Nobody owns the analytics

Another common problem is when nobody is at the finish line to measure what’s happening. For example, the marketing team for a fast-food restaurant chain used AI to quickly launch a campaign for a back-to-school offering of free food for high school seniors every Friday in September. They were really proud of the quick delivery, but when leaders asked how many redemptions the campaign created, the team stared blankly.

Who owns this job? They had no idea where the data lived, let alone how to capture it.  The result: faster wrong answers, not faster right ones.

Seeing this process gap was just what the team needed to tighten their operating model. Now they have a dedicated marketing analytics team member, which gives them regular access to their data and enables more informed, data-driven decisions.

What AI’s exposure means for marketers

Sorry if this article has left you feeling like your team’s process is standing there naked and fully exposed. The good news is you can fix it. It’s time for marketing organizations, yours included, to move beyond “Which AI tool should we use?” to “Which processes need fixing?”

If you’re experiencing any of these common AI exposure problems, here’s what you can do right away: Map your current process from end-to-end and find where the bottlenecks currently live. Then work as a team to develop solutions to fix them.

AI isn’t just another tech tool; it exposes problems related to the organization and its people. Don’t roll out another tool without considering what currently works and what doesn’t in your process.

Everything enterprise teams need to grow visibility across search and AI.