How I Became "The AI Guy" (And How You Can Too)
You don't become "the AI person" at your company by accident.
You become it by being useful.
I'm a Director of AI Enablement now, but that title came from years of building things, solving problems, and (eventually) having people notice. My life prompts resulted in an output of being an AI director. See what I did there? Hey, I'm a dad. We have bad humor.
Fair warning: I am not an engineer. That's a totally different aspect of AI. I am a business solutionist. I identify business problems and architect ways for technology to solve them, most of the time with AI, but not always. There is still plenty of room for non-AI automation and process improvements.
Here's what actually got me here, and how you can do the same.
AI Is a Combination of Various Skills
I have given multiple talks to companies and teams about how to use AI. Recently they have become somewhat basic. This is natural as I have become more comfortable with the space. I mention it here because too many people online complicate AI.
There are a few key things to remember when you want to work with AI. Once you have those down, it's repetitions and continual learning that take you to the next level.
First, you need a paid version of Claude, GPT, Gemini, or Copilot. Paid versions have better features and, if you're in corporate, they're more secure. That's table stakes for everything that follows.
With that in mind, becoming skilled in AI comes down to a few key areas.
Data Analytics
It's shocking how many degrees do not teach students how to analyze data. Really shocking.
I was fortunate to attend a graduate program that emphasized data analytics even though I was in a "soft science": anthropology. People would ask me what I did as an anthropologist and my answer was always the same. Collect data, analyze data, write about data, and present data. That's it.
That ability to analyze data has served me well for the past 15+ years as I worked in marketing with an emphasis on data analytics.
With AI, data is the most essential ingredient to building business solutions. AI works with real data. If you are relying on AI to pull internet data for a major business decision, you might as well have Reddit run your business for you. Yikes.
What does that mean for you? You need to get better at understanding data. Really understanding data, not observing data.
Here is what most people do regardless of their field of work. They say they know data, but what they are really doing is reporting on observational data. This is when someone shows a number but can't explain anything about the number. Website traffic is up and they cheer. Great, why? The numbers are effectively meaningless without an underlying understanding of the forces driving the data.
Flip this back to AI. You will need to feed your AI platform real data, but more than that, you will need to explain the data to the AI. Once you can do that, AI can do incredible things that would normally take humans dozens or hundreds of hours to execute.
Here's the difference:
Bad prompt: "AI, give me a marketing plan for Q2."
Good prompt: "I've uploaded our Q1 performance data, competitor analysis, and budget constraints. Analyze this data and recommend Q2 priorities based on what performed best."
The first prompt asks AI to invent information. The second gives AI real data to analyze. One produces generic advice. The other produces something you can actually use.
I have built multiple agents that can break down data faster and more reliably than I ever could. They can even run analysis that was only a dream a few years ago, like identifying attributions in marketing campaigns across channels that typically don't connect to each other.
I know that was a lot about data, but it's probably the most important element of AI. You need to use real data. Most of the AI haters you see are operating from a framework where they think AI should know everything because it's "intelligent." They don't understand how AI works. Don't argue with them. It's not worth it. You want to be the AI person. You don't have time for them.
Focus on becoming better at data analytics and using real data. And don't stress too much.
Once you get the real data, you can simply ask AI some basic questions:
How would you analyze the data?
What would you look for in the data set?
What trends are there in the data?
What anomalies are in the data?
The great thing about AI is that it doesn't even need to be data in the form of spreadsheets. You can feed AI text documents, emails, PowerPoints, and other items you and I routinely work with.
Become a Teacher
Let's reconvene. What do we need so far?
A paid account
Real data
Now it's time to become a teacher.
Here's another advantage I have. For most of my twenties and thirties I was a teacher in one form or another. I was a personal trainer, then taught at the University of Houston, then was an intern leader at a non-profit, and finally led a global team in my last role. These jobs developed my ability to teach.
AI is an ideal student, but the platforms require extensive teaching to work the way you want. Anything from a single prompted output to an agentic workflow requires detailed instructions.
Many of my prompts easily hit multiple pages of input and I rarely, if ever, rely on a single output. There is a discussion with AI until we reach a final approved output.
Similar to data analytics, it's frightening how many professionals are never in a role where they have to teach others. Don't take my word for it, but there is plenty of research to indicate mastery occurs after you can teach a skill to someone else. Think about that for a minute. You really haven't mastered your craft until you can elegantly teach it to someone else.
Hate to break it to people, but that's why most people are ineffective with AI. They can't teach a human, let alone an advanced computer system, how to execute a task.
If this resonates with you, here are some simple ways to become a better teacher.
Take a task you do frequently at work and open a Word document.
Type every step it takes to execute your task. I mean everything: opening a web browser, clicking a folder, opening a file, whatever it is you do. Thinking through each step helps you break down the tasks, and writing it down codifies it in (you got it) real data.
Now you have a fully formed list or prompt that you can use if you want to try and semi-automate a task.
The more you break down tasks, the better you will get at teaching. You don't even need to write it out. You can use voice to text if that helps. The point is you need to become a teacher if you want to become the AI person at your job.
One of the great things about AI is that platforms like Claude are always there to listen. You can develop your teaching skills with AI by continually teaching it things and then asking some questions:
Was I able to teach you [insert what you were trying to teach]?
Did my teaching make sense?
How could I improve my teaching?
How This Actually Played Out
Let me give you a concrete example of how these skills compound.
Early in my AI journey, my team was spending 40 hours a month on market share reporting. Manual data pulls, spreadsheet cleanup, building the same PowerPoint deck over and over. The process worked, but barely.
We mapped the workflow, identified where the bottlenecks were, and realized most of the time was spent on data wrangling, not analysis. So we rebuilt it. BigQuery for storage. Python for cleaning. Dashboards for visualization.
40 hours became 2.
That project got noticed. Then someone asked if I could help with their reporting. Then another team. Then I was leading AI training sessions. Then I was advising on enterprise AI strategy.
I didn't pitch myself as "the AI guy." I just kept solving problems and showing results. The title followed the work.
That's the pattern. Build something useful. Let people see the results. Repeat.
Last Thing: Be Objective and Identify Impact
If you are keeping a running list, to become the "AI person" at your job you need:
A paid account
Real data
The ability to teach
Here is the most difficult item: you need to be objective and identify impact in your role.
We have all experienced corporate theater where we are sitting and thinking:
This meeting could have been an email
This PPT is pointless as no decisions will be made
I'm giving a presentation and people are just completing emails while I talk
You know the feeling. It's soulless and demotivating to realize you are spending your precious limited time on something that really doesn't matter.
How does this play into AI? We will get there, but first: for your work, think about how you are impacting the traditional items that improve businesses. Are you:
Increasing revenue (this doesn't have to be sales; you could be designing a product, working in marketing, and many other forms of revenue generation)
Improving brand awareness (are your actions increasing how outsiders view your company? This could be customer support, marketing, event planning, and more)
Gaining market share (product management and logistics are key movers here)
Decreasing costs (every department can improve efficiencies to help reduce cost)
Those are the key movers for any business, and your daily activities should roll up to one or multiple of those areas.
Now, with AI, here's where I have seen a lot of issues. People use AI like a party trick. They have AI write emails as a pirate (yes, I have seen this). They have AI do work that amounts to no impact. They have AI build marketing plans with zero data or teaching. And on it goes.
You need to take a moment and objectively determine if what you are doing is creating business impact.
I ran my own business for five years, and this was a constant conversation I would have with clients. I would have to point out areas of their business that simply existed but did not create value. It's uncomfortable and something many people refuse to do. Well, that's a problem.
Put on the big girl and big boy pants and be honest. Is your work making an impact? Identifying the top impactful actions will lead to AI projects, as you will be able to amplify those activities.
Here Is How You Do It
Determine one redundant task you do weekly, monthly, or quarterly that adds value. This could be a report, marketing campaign, sales update, sales outreach, or a thousand other items.
Find your previous three outputs. A report is an easy example. In this scenario you would find your previous three reports.
Feed those three reports to your paid AI platform. This is your real data.
Explain to your AI platform the goal of the report, stakeholders, how it's used, the desired output, that you have given the previous three versions, and how you want to improve it. This is your teaching.
The next time the report is due, split the work between you and AI. Go prep the report like you normally would, but have AI run the analysis and see if you agree with the output. For the first time you will likely need to tweak things, but you will get to a desired output.
You just ran your first end-to-end AI project.
Tell your manager and your skip-level manager how you used AI to accomplish the task.
A note on that last point: this is a double-edged sword. On one hand, you will be seen as an "AI person." On the other, you might be given more projects. Be mindful of how you share your work.
The Compound Effect
Once you do the above once, twice, three times, you will start to get in a rhythm, and others will notice. It's not a matter of if, but when. You know how it is. Corporate folks observe and talk. It only takes a few AI wins to be seen as an AI person.
The reason there are so few is that they don't follow the above. They don't use real data. They don't take time to teach AI. They don't use AI for impactful work.
That doesn't have to be you.
Use real data. Become a teacher. Be objective about impact.
Nobody becomes the AI person by reading about AI. They become it by building things that work.
Start with one project. Show the results. The rest follows.
