17 things I learned talking to Marketers about AI for 3 months

I've spoken to everyone from CMOs to new hires. Some in-house, some at agencies, and others at AI-first orgs.

Maybe 10% of teams felt like they had a handle on it (even fewer had a strategy behind it).

What's happening inside companies sounds nothing like what's happening online. Here's the truth:

Everyone had access to a chatbot. Mostly Gemini, ChatGPT, and Copilot (few mentioned Claude). There were also a good amount of companies that built their own chat interfaces, likely built on top of one of the major platforms, but with tighter privacy and fewer capabilities. For most, it seems like only half (or less) of team members use the chatbots they have access to. 

2. There's a lot of shadow AI happening. 

It's very common for people to be paying for unapproved tools out of pocket. The ones I heard most were Gamma (slides), Granola or Tactiq (meeting transcriptions), and Claude or ChatGPT (as a second chatbot outside of their org approved one). Many feel like they can't openly tell others what they're using.

3. No one feels adequately trained or supported.

A good amount of teams had zero training. For many there are resource hubs available or optional training, but unless it's mandated they're not likely to check any of it out. The ones that did have training were in the format of compliance videos (watch to check a box, not applicable).

4. Almost nobody is excited, I sensed dread. 

AI is just another thing on people's plates and one that has a lot of unknowns. There is general fear around it taking people's jobs, but more long-term than near-term. People know they should be doing more, but are too busy and don't know where to start.

5. Teams are left to figure it out on their own. 

There's a lot of conversation and activity, but no clear leader. Team members are waiting for managers. Managers and leadership are waiting for IT or L&D. IT and L&D are figuring it out too and are lacking a view into other team's needs. There's a lot of frustration about there being no helpful guidance or strategy from the top. AI is everyone's job, which also makes it nobody's. That's a huge reason for the spinning and stalling.

6. People are protective of their stuff. 

On a few teams, I heard about people gatekeeping good prompts or workflows they created. A way to stand out or to protect themselves as fear takes over about AI downsizing.

7. AI was in about 50% of performance reviews.

Roughly half the people I talked to mentioned AI is being added to their annual reviews, but specifics on it are vague. Of those who had AI goals last year, some admitted to fudging their scoring because there was no clarity behind the rollout. 

8. AI can be too good at brainstorming.

In a few cases, I heard about AI surfacing big ideas like campaign ideas, creative pitches, and event concepts that were oversold. Ideas were presented and well received, but when teams went to execute them, they realized they couldn't bring them to life (too much budget, no capacity, not possible).

9. The slop problem is less about receiving slop, and more about the fear of sending it. 

I expected to hear a lot of people annoyed by the amount of slop they're getting, but it was actually the opposite. A lot of people couldn't pinpoint being sent slop. Teams are training each other to be cautious (in some cases it seemed overly cautious). 

10. Everyone is namedropping agents, but hardly anyone is building them. 

Agents is a major buzzword. The word was mentioned in most of my conversations, but the vocab is way ahead of the practice. Few people actually understand how agents could fit into their work and even fewer have made real progress creating them.

11. Marketers who could actually use agents are being left out of designing them. 

Larger strategy and building is happening among Leadership, IT, and Ops. This is a big miss. Redesigning processes around AI needs guidance from the people in the work. 

12. Marketers are defaulting to vendors for AI growth efforts. 

Most are on the hook for showing what their team is doing with AI, but haven't been given proper time or support to do it. So adding AI vendor capabilities is the easiest win. Many are exploring visual prototyping, Meta offerings, and creative versioning. There's also skepticism from vendors slapping AI onto everything without being able to explain what they're selling.

13. Creative teams are dragging their heels more than others. 

Bigger picture, they have good reasons like protecting their brand and public backlash. The downside is that they're missing out on faster and lower cost concepting, versioning, and testing. It's also spreading hesitancy into adjacent departments. 

14. People aren't sure if AI actually saves them time.  

I asked where AI was most valuable, which was a surprisingly difficult question for people to answer. A lot of people are spinning with it. The output might be a little better, but it took the same if not more time, and in many cases caused more frustration. This ends up being a cycle, with more resistance to trying it on new projects or work with tight deadlines.  

15. No one named the gap between what AI can do vs what they're using it for.

Most information out there is theory. It's not practical or tied to someone's responsibilities, so people default to basic use cases like brainstorming, drafting emails and outlines, and asking questions. People know they're underusing AI, but they can't name what they aren't doing or want to do more of. Hardly anyone named specific processes their team is using AI for either. It's ad hoc, not structured across team members. An end-to-end process, even patched together with structured prompts or defined steps, is rare.

16. No one brought up ROI unprompted. 

In studies, ROI is often brought up as a huge barrier for investment and adoption. I almost never heard it in my conversations. The people who care about ROI are not the ones figuring out how to use the tools. In most cases, their teams are way too premature to be thinking about delivering ROI. At this point most teams need to understand fundamentals to be able to design anything that has a chance at being measurable.

17. No one feels confident they're ahead of the curve. 

Even the teams clearly further along than others were surprised when I let them know my thoughts. There's no shared sense of what good AI strategy and adoption looks like right now. It's usually the teams that don't feel behind with AI that I'm more concerned by.

The 10% figuring this out should scare you. They're getting more capable each week.

Just check out how quickly AI labs are able to spin up new offerings. That speed will hit your industry too, whether it's new businesses fully built around AI or small pods inside your own org quietly pulling ahead.

Competition and expectations will keep getting louder, which is exactly why your org has to move now.

I built Chasing Next to help teams get into the 10%. If you read this and recognized your team in too many of these, that's your sign to press the gas. 

Reply "AHEAD" and I'll send you a personalized link showing how we can get your team there.