I am fascinated by the future. I also am worried about it.
So I’ve been writing down a ton of thoughts about the 5 stages of digital marketing career maturity as it relates to AI in marketing, trying my best to give sights and sounds for people to assess themselves.
Almost 5,000 people have given our AI marketing maturity assessment a whirl, with 100+ ratings holding at 4-5 stars. Get a roadmap telling you where you are, and where you might need to go to move to the next level of maturity.
2 years ago I wrote this piece on the AI survival guide to your marketing career - then it got old, the world was moving fast, so I wrote this post about 8 months ago, and now I’m writing this one.
My goal is to be maximally transparent about my thinking with my team, and I figured why not post it publicly for others to share their thoughts, challenges, etc.
I caught myself doing some low level stage 2 shit
I am supposed to be at least a 4. I figured lets be honest about it, make this a post on how I cuaght myself, and then went from stage 2 to stage 5 in a weekend - using this one example of an industry update. Catching yourself is 90% of the battle.
Here's my attempt to map that spectrum honestly, using something that happened to me this past weekend as the through-line.
Why This AI Marketing Maturity Framework Exists
There's a huge chasm between "I asked ChatGPT to write an email or summarize a post" and "I built infrastructure that automatically surfaces insights across my entire client portfolio and compounds my own thinking over years." This is the difference between stage 2 and stage 5. What used to be stage 5, ooooh it summarizes using ChatGPT 3.5 is now stage 2 in a few years.
The Google Marketing Live post I read this weekend is a great example happens DAILY in our industry, it is why we’re all overwhelmed. A major announcement drops, a new study, an algorithm change, a layout change, a penalty, and the question isn't whether you read it or not anymore. The questions I’m starting to ask is what did I do after I read that post I just read?
How to use this...just read it and be honest with yourself on where you are. You can choose what to do with that afterwards.
Stage 1: Overwhelmed, Falling Behind
Ask yourself this 1 question: What thing have my leaders done, related to AI in the last 2-3 months, that is the result of me being on their team?
Stage 1 marketers, when it comes to AI are stuck in a bad cycle, which means as a leader they are the anchor around your neck, taking the ship and crew down with them.
They'll tell you 3 weeks from now that it is on their “to-dos” to read the new posts from the Google Marketing Live announcements. They'll tell themselves that there’s no way to get caught up on it all, they're too busy. Sadly, the reason why they don’t have time to get caught up on the changes in AI, is because they haven't used AI to free up time on basic deliverables.
It's a self fulfilling prophecy.
-
The more new stuff launches
-
The more overwhelmed they get
-
The more overwhelmed they get
-
The less capacity they have to step back and ask whether there's a better way
- The more they do things the old way while getting more pressure
The result: “going heads down and grinding” - which isn’t good for them or your business.
Complaining isn’t the problem, doing nothing with your complaints, is.
Stage 1 AI marketers complain a lot, and don’t realize that recording their complaints is the actual way out.
Here's the other thing about Stage 1 folks, they complain mostly to other Stage 1 people. Cause stage 2-3-4-5’s realize they got the same 24 hours in a day as the stage 1’s and that stage 1 complainer is wasting their time.
Get out of stage 1 with one simple thing, a voice recording of your complaint, struggle, issue.
Record whatever you are complaining about and put it into whatever AI you can get your hands on, tell it you want to speed up 25% of this and to give you 3 very different ways to do it.
That's the move Stage 1 never makes. Complain with the intent to improve. I complain all day, and I turn to claude to help me.
It's not hard you are already looking at the work and complaining about it, you already are going heads down and banging it out, record your complaint and record your work while you are doing it.
Send it to Claude. Ask how to cut it in half.
Heck at our company with our 150 skills and our company wide MCP with all client performance data, you could prompt your complaint, your work, and simply ask “what skills exist already in our company to help me with this” or “using our MCP how could you help me?”
They will never catch up, because they won’t use the tools and training in front of them.
Stage 1’s will feel the pressure of a shrinking portfolio of clients willing to accept low AI team members.
The way I described it to my team was this in an all hands 6+ months ago the pool of clients who won’t let you use AI isn’t growing, it is ONLY shrinking. Your career prospects will shrink with them too. Each new post is something to bookmark with good intentions, but never read, and never act on. The world is changing too fast. Sure there’s cars, but they always walked, so they can’t fathom a world where you can get 100 miles in a day - and sure you might have to slow down a bit, learn a new skill (driving), get a license, but once you go through that you are on the right track.
Wanna get out of Stage 1?
Stop complaining to people who like receiving complaints.
Start asking your AI for an alternative perspective, ask for coaching, ask for your blind spots, then ask for solutions, see what comes back.
Stage 2: One Time vs Portfolio approach
Stage 2 marketers read the Google article, it’s not easy but they are keeping up, barely. They are who you want to spend your training time on, cause they are the ones most likely participating a bit in your company chats and what not, they are actively trying to learn but just like stage 1 the waves are crashing over their heads and they are drowning some days.
They might summarize with AI and get a solid grasp of what changed in the article. They're engaged. They're trying. They understand that AI is something they should be using and they see signs of light.
But then they'll do exactly what I first did in my example above, read this post, summarize it and maybe ask about a client that is on my mind lately, using AI.
How might this new AI-powered shopping ad format affect client X. They'll go back and forth with the AI, build a real thoughtful insight for that one client.
The Stage 2 teammate is genuinely adding value in that moment. They're using the tool. They're applying it to real work.
The problem is in this AI world, anyone in marketing and in an agency sees the writing on the walls, somehow this company has to pay for this BIG investment in AI, and at some point parts of my job will be automated, leaving me with MORE clients to justify my salary.
That is a real squeeze. They have 5 clients today, but in a world where assembling reports and what not is going away, they see a world where they might have 8 someday, and right now they assume they’d have to do this 8x to help every client, and that just might not fit into their time horizon.
In my example above I saw the post, summarized it, talked to claude a bit, but I did it for one client. I legit think I am a stage 4-5, and yet in that moment I was a stage 2. :) It happens to all of us, sometimes we all need someone to say “did you google that?”
I caught myself - I realized I was going to spend half a day doing manually what a better workflow could do in five minutes.
Now building that workflow is going to take 5x longer TODAY than if I went “heads down” and did it manually. But your stage 2.5’s growing to stage 3’s will catch this and seek out help / try themselves.
Today’s "LONGER work" to harden the AI output, is tomorrow’s "never work".
I’m automating the assembly of work everywhere I can. Leaving me with as much thinking and hands on consulting work (across more clients, but doing higher value work),
The Stage 2 trap for marketers using AI is that it feels like progress because it is AI progress. You're engaging with the tools, you're applying them to real problems, you're not drowning. But you're starting from scratch, maybe you have a prompt library (remember when people thought that was going to be huge), but a prompt library of saved prompts exposes you as stage 2 the world has moved WAAAAY past that. IMHO you are tomorrow’s stage 1 or a burnout victim - if you don’t keep up.
Your system don’t scale.
Stage 2 employees are often the ones who will tell you they "use AI every day" because they do, they use it to draft emails, summarize articles, prep for calls, and act as a thought partner. That's real usage. What makes a stage 2 a stage 2 is that they haven't asked the next question, which is: how do I make this work across everything I do, not just the thing I'm doing right now? They're getting efficiency gains on individual tasks to move into stage 3, you’ll find they start building things for their entire portfolio, and they share internally because they realize that every time they spend MORE time today to build something that scales across their portfolio, they are only a slack thread away from showing 50 other people that they can do the same.
The move out of Stage 2 isn't working harder or even working smarter on the tasks in front of you. It's changing the unit of work from "one client" to "my entire portfolio."
FYI one of my favorite bangers this year came from Allison worms, we were all internally talking about how we run out of tokens sometimes, she was like I kick off a few tasks on Sundays so I don’t burn up my tokens on monday with all my ideas. Smmaaaaart!!!! This is what it looks like to move into stage 3, find a way, be efficient with tokens.
Stage 3: Working Across a Portfolio
Stage 3 is where leverage starts to appear - this isn’t quite yet orchestration, but we are on our way to it.
A Stage 3 employee sees the post from Google Marketing Live and says - Based on all my client calls and all their account data, which clients would this update impact the most?”
Now you might be thinking, holy smokes calls take up a LOT of my context window, well yes it does young padawan, that is why we built a whole meta data tagging layer (Thanks Lauren K) on top of every one of our client calls, keeping us from having to read entire transcripts to get the gist of what happened on the call, custom tuned to what we look for in calls - for now, just get started with that you got.
If you are thinking wait, client performance data, but how do I keep it from hallucinating? We also built detailed schema and our tools fire off SQL to pull answers, so our agents are pretty resilient to data hallucinations because they are built and tested to only write SQL - you can do this too. I’ve shared my examples of how I do this for GA4 with my project instructions.
We had to build this stuff at scale, my point here is to show you what's the next level - you have to do stage 2 work for one client before you get to scale, just keep this in mind, when you go to scale you will run into new problems, but those problems are worth solving to eventually get past stage 2-3, the obstacle is the way!
The operating system of stage 3’s AI forward marketers is different lets look at “what gets done with transcripts” for all three stages, each builds on the next.
Stage 1 - great this gives me summaries, notes, and next steps. Perfect.
Stage 2 - uses stage 1 + drops in a raw transcript for coaching, to ask a question about a post, get ideas and feedback
Stage 3 - uses the ideas from stage 1 and 2 + has a zapier workflow, a claude co-work workflow, doesn’t matter but they see the transcript as something to scale, they have automated workflows to turn that same transcript (raw materials) that the exact same person at stage 1 got - but they spun it into gold that helps them retain clients better, get better at their job, and show their manager that they are improving.
A stage 3 takes that same transcript and makes gold from it.
- Coaching after every call, so when they follow up, they can give a “hey I missed this”:
- Automated comparison of client calls to their last employee review so they can show their boss the improvements they are making
- Automated trending of client sentiment, etc etc
At Seer, we've made this accessible for our whole team. We have an MCP (Thank you Ethan + team) that connects Seer Signals directly to paid search, SEO, and Generative Engine Optimization data across every client inside Claude.
That means anyone on the team can take a blog post from anywhere Google Marketing Live or something they just read, drop it into Claude, and ask: "Based on my account performance, how big a deal is this and what metrics would be my canary in the coal mine that would indicate that this is a problem or an opportunity?"
This scales portfolio wide. This is stage 3. They will also put in the LONG work today to minimize hallucinations, so they never have to do it again. Our scaled client transcripts are launching this week, allowing them to also look at what clients say on calls at scale.
When I asked which clients were most likely impacted by the new Google Ads formats, I got back a prioritized list:
- A client in the pet food space was flagged, why cause it looks like chewy was mentioned in the post
- My travel clients got flagged because the article mentioned Booking.com and Expedia in the post
- My higher Ed clients got flagged because a higher-ed example was given in the post, showing how people wouldn’t even have to the university’s web site to complete an application right there inside AI mode with an ad unit.
I gotta be honest, hundreds of hours of building out our MCP, and the team having the vision to build it was a major unlock here, that enables all this to happen at enterprise scale.
Stage 3 people are starting to work at portfolio scale not company scale yet, give them more tokens, and nudge them to “google it” for prompts like “how could I alert every team member on every project” and see what comes back. They've built a workflow that makes them more effective, your job as a leader is to help them see this as something to scale for everyone, something to share internally.
Stage 4: We build infrastructure
Here's where I actually ended up on Saturday.
After I got my prioritized client list, I didn't stop. I thought: this is going to happen again and again and again.
Google will announce something else. And next time, I want this to just... work. I don't want to have to remember to check these sites, I want the insights dropped off and thank god with our skills and MCP I got to the tools to do it.
I don't want it to depend on me having the presence of mind to drop in the right URL at the right time. I want a system. It only took this prompt to turn a one time job into a scalable solution company wide.
Find all of the official Google blogs RSS feeds, I want you to build out a skill that takes these RSS feeds, finds out all the changes, runs a prioritization of how important of a deal it is, and then if it is important enough runs the post against our MCP looking for past, current, and future clients we can update.
Again the MCP to the rescue, it syncs with our CRM, so I can even look at people in the pipeline whom this might impact.
So I asked Claude to build a skill: pull RSS feeds from major Google properties, detect relevant changes, run them against our MCP data, identify which clients and prospects are affected, and surface the ones worth a conversation, with human review built into the loop. I got real pushback on parts of it from Claude. Some of my ideas were wrong. I incorporated the feedback. By the end, I had a documented skill I can now implement and share. Now top be honest I am still in eval mode, but it will be shared when I get it done.
Stage 4 AI Marketers isn't just about personal efficiency. Stage 4 people lift stage 2’s and 3’s - give them infrastructure they can reuse as they see fit. Especially your stage 2.5’s and 3’s are leaned in enough to take advantage of it. They share the skills they build. They bring the team up with them.
The Stage 4 move is: I did this manually, I learned something, I encoded it into infrastructure, and I made that infrastructure available to others. That's a compounding return. Every time a Google announcement drops, every time a platform changes, the system is already running. The team already has the tool. Nobody has to start from scratch.
What is stage 4 missing?
What Stage 4 often misses, though, is something subtler: the compounding of their own thinking. They've automated the intake, the pushback, the building, the scale. What could be missing?
They haven't systematically captured what they believe about what the system surfaces.
I think a Stage 4 marketer is really great at making this scale, but what they probably haven't done is library their thoughts. So there's a chance that three or four years from now, they've got these things at scale, but they haven't inventoried what they thought about all of these things.
They've got the systems. Now its time to add their unique value.
So far everything I’ve articulated is eventually automatable. EVERYTHING, it could be 4-5 years from now, it could be 10, I don’t know but I do know it is NOT a moat, it is moving 0’s and 1’s.
Lets say all the stage 3’s and 4’s have built all the dope stuff, and I see a real role for people who constantly read about their industry and manage and improve a set of skills / automations, and MCPs with the newest information and intelligence, the newest data sources, APIs, etc.
Stage 5: The Orchestration
Stage 5 does everything Stage 4 does and adds two things, 1 they realize what parts of what they do today are automatable tomorrow and 2 they understand that there's no moat in automation and they go looking for where else they can create value.
Every time I read something that impacts how I think, I record it. Voice memo, Readwise highlight, etc, the format matters less than the habit.
The point is that I'm building a database of my own thinking that compounds over time FYI those of us who highlight on the kindle ar at an advantage.
You are consuming my digital exhaust right now, this is my 3rd or 4th post on AI marketing maturity on the Seer blog, I am documenting my thinking through the act of blogging, posting on linkedin and YouTube, reading books and articles about my industry.
Every one of these is digital and stored, blog is easy, scrape all my posts and store them, linkedin, export all my posts and post performance, or use Phantombuster to pull all my comments and posts, take all my youtube videos and take the transcripts and store them, take all my presentations and store them, and use the readwise MCP to index all my thinking over the last 13 years. This is building a MOAT that brings me back my own ideas at the right time for the right clients. Letting me stretch across MANY more clients. This is where stage 4 and stage 5 AI marketers start scaling, I am able to use my brain to help 10x more clients than ever.
Stage 1 and stage 2 work is drying up
Cause higher stages are scaling insights
This dries up work from Stage 1’s and 2’s - what would trickle down before is gone, why would I:
-
Have the stimulus (reading the post, now I got ideas)
-
Post it internally
-
Hope you read it (oh you are OOTO, I’ll wait)
-
Read your response 2 weeks later after you caught up.
When now I can
-
Summarize the Google Marketing Live post (stage 1)
-
Banter with AI about my thoughts & applying it to one client working out the kinks (stage 2)
-
Apply it to a portfolio of clients (stage 3)
-
Construct the infrastructure to apply to all clients and alerting team members (stage 4)
Ok Wil, now every person in your company has the same information, you are more costly than they are, so why wouldn't I use them over you to save cost?
Stage 5 is where my value is created, all the books I've read, the 7000+ highlights, the 1000s of presentations, internal videos, and 20,000+ articles I've read since I got readwise. And the fact that almost every weekend I read 30-50 articles to stay ahead of my automated tools and apply my thinking.
I'm using me vs my team here as an example, but this is agency 1 vs agency 2. At some point either 6 months or 6 years from now what is above will be normal.
What won't change the limited attention and real estate for marketers, so being one of 10,000 good marketers doens't matter if you can't get into the limited real estate where the organic and paid answers are.
Those fractions of a percent of what we do differently will hopefully be EVERYTHING.
Applying all my thoughts, voice notes, highlights, presentations, etc to ask - what have I read or thought that conflicts with or doubles down on this post from Google Marketing live. Then I slide off into the DMs of other people who make me think, keep me sharp, and give me perspectives I wouldn’t have otherwise had, and wrap it up.
Helping my clients, with my brain at scale, as I am thinking it is now possible and happening.
The other things about being a stage 5 is that now I am a few loom videos away from messaging my clients to say hey here is what happened, and here is what I am thinking, at scale. The more I automate the other tasks, the more time I have to read and shape my own thinking, and the more time I have for clients and for my team to help them use my files and build on top of my infrastructure.
Be a stage 5 who hardens the work of prototyper style people like me.
One thing I haven’t mentioned is that often when I build at a stage 5 level I like to get it to my engineering team to harden it, now is when I truly need a partner to scale it.
I have proactively recorded all of my thoughts about this change. By doing that, I'm adding to my database the thoughts every day when I read a post. If I have a unique thought where it makes me think in a new way, I record that new thinking. And I store that new thinking so now I've got my thoughts stored in an ever-expanding database.
I'm bringing years of accumulated, retrievable perspective to it in real time. I'm not going into Drive and trying to search or going into Obsidian and trying to search.
The automation can tell you which clients to call. Only you can tell you what you actually believe about the change, why it matters, and what the client needs to hear from you specifically. Stage 5 has been building that second layer for years - AI is just the key to unlock all that value.
Second: stage 5 marketers orchestrate tools and build what doesn't exist yet.
"Once things are served up to them on a platter from stage three or four, they then start seeking other places they can scale this. So for me, as a stage five, I'm now orchestrating several separate tools that I've built to go out and do research for me."
When I read the Google Ads announcement, I wasn't just thinking about what my existing systems could tell me.
I was thinking about what I'd need to build to stay ahead of this category of change entirely. A stage five person would say, I should build a screen grab tool that will screen grab a couple of prompts in my industry every day to see if this type of ad is starting to show up. And then they're going to share that with the entire team.
Stage 5 people aren't just using the tooling that exists in the organization they are building prototypes of new tools, checking the value with clients, then when value is created for clients seeking help hardening it. They identify the gaps the current infrastructure can't reach and they fill them.
Who are you lifting up, and can you measure it?
- Stage 1 brings up no one, no teammates, no clients, not themselves / their careers
- Stage 2 brings up a few clients, if they share internally, they might inspire others to do the same
- Stage 3 brings up the whole portfolio of clients, they bring up the company too
- Stage 4 brings up teams team
- Stage 5 lifts the team, compounds their own thinking, and expands the ceiling for everyone
Here is a smaple prompt:
Scan my emails, slack, and calendar looking for meetings I have, ideas I am dropping that show me putting my ideas out there, did people respond saying they were going to test it? Is that work today inside of any of our company skills, custom GPTs, etc.
One More Thing About the Google Ads Example
One insight that came out of my Saturday session that I'd never have gotten from just reading the article:
For clients running AI ads (now we’re seeing them in ChatGPT and Gemini) what if the questions they're asking and pulling from your website are getting things wrong, causing you to get leads filled out, which would then train the algorithm to do more bidding because this is working well, but then those leads aren't converting on the back end.
The fix? Track branded prompts across engines. Are your ads saying we’re great and the organic answers are saying nope “don’t use this company for your use case?”.
That's the kind of insight that only comes from using AI as a thought partner, not just a summarizer. That insights only comes from collaborting with other humans who make you think, hard and collaborate with you.
And it only compounds if you capture it somewhere - which is exactly what separates Stage 4 from Stage 5. Build that network of people who make you think.
The gap between stage 1 and 2 and 2 and 3 is widening fast.
The question isn't whether you're using AI anymore. The question is: are you building anything? Are you capturing your own thinking in a way that compounds? Are you expanding what's possible for the people around you?
That's the difference between keeping up and pulling ahead.
We're hiring hungry stage 2's who want to be around others trying to move up the AI marketing maturity ladder.
If you are stage 2 or 3, and think this kind of thinking is the kind of people you want to be around… we got some open roles - I got a lot to learn and a lot to share, if you build the systems to keep up.
Crystal Oneill (our co-CEO) built shared the prompts on how our agents help you put a problem into Ninjacat and find out who in the company is tackling that problem.
Jordan Strauss built a skill that helps you build skills :) The idea is when you are working on things in claude, wouldn't it be nice to have it say "this could be a skill" - let me check our existing 150 skills and see if your changes can help everyone?
Ethan Lyon helped me understand MCP over a year ago and got me up on it. Early now he and our team are building it out internally (and are hiring).
This is what it feels like to try to build the infrastruture of a stage 5 agency.
Wil Reynolds
CEO & Vice President
