Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS
What’s up everyone, today we have the pleasure of sitting down with Tata Maytesyan, founder and CEO of Grow Global Tech, where she builds AI marketing systems for tech scale-ups.
Summary: Tata breaks down why the best AI automation targets are the boring, repeatable tasks nobody talks about on LinkedIn, and why the specialist-to-generalist shift in marketing is already happening whether your org chart reflects it or not. She also gets direct about the 10,000-hour threshold for building genuine competence across domains, and the self-preservation fear she hears from leaders every week. If you have ever wondered whether you are building your career around the right foundations, this episode is worth your full attention.
In this Episode…
- Every Marketing AI Project Should Start With a Whiteboard
- Why Deep Generalists Outperform Channel Specialists in Marketing
- Why Marketing Org Charts Are Not Getting Flatter
- Most Marketing Teams Buy AI Tools That Nobody Actually Gets to Use
- How to Keep Your Marketing Job When AI Is Automating More of It
- How to Figure Out Which Parts of Your Job Are Worth Keeping
Recommended Martech Tools and Agencies 🛠️
We only partner with products and agencies that are chosen and vetted by us. If you’re interested in partnering, reach out here.
🔄 GrowthLoop: The agentic, composable CDP that drives compound growth by uniting your cloud data + AI into one marketing engine.
🐾 AttributionApp.com: Multi-touch attribution software that gives you auditable, full-funnel clarity across every channel — find out your “true” CAC.
📧 MoEngage: Customer engagement platform that executes cross-channel campaigns and automates personalized experiences based on behavior.
🎨 Knak: Go from idea to on-brand email and landing pages in minutes, using AI where it actually matters.
About Tata Maytesyan
Tata Maytesyan is the founder and CEO of Grow Global Tech, where she builds AI-powered marketing systems for tech scale-ups and runs a hands-on AI bootcamp for marketers on Maven. She spent 15+ years leading growth inside Nike, Deloitte, and Picsart, including a stint as Head of Product Strategy and Operations for Picsart’s content and AI division, a platform with over 100 million monthly active users. She has since advised more than 40 companies across 12 countries on go-to-market strategy and AI adoption, and consults primarily with CMOs and CEOs at companies between a few million and $200 million in annual revenue.
Every Marketing AI Project Should Start With a Whiteboard
The wrong starting point for AI adoption in marketing is inspiration. Most marketers scroll LinkedIn for jaw-dropping use cases: ad creative generated at scale, competitive analysis in 10 minutes, entire campaign briefs written by agents. It looks impressive. It’s also almost never applicable to your specific job on any given Tuesday. Tata has spent years watching this pattern play out with consulting clients and bootcamp students. Her fix is deliberately boring.
At the start of every engagement, she asks everyone in the room to close their AI tools. Then she opens Miro and maps how the team actually works. From there, 3 questions run against every process on the board: how often the task repeats, how acceptable an imperfect output would be, and whether it’s something you actually enjoy doing.
“I always think and say workflow first, automation or AI second.”
Those 3 questions quietly eliminate most of what people think they want to automate. Frequency kills off exciting-but-rare workflows not worth touching. Risk tolerance separates contexts where imperfect output is acceptable (most content tasks) from those where it isn’t. Tata advises a healthcare client where certain work is patient-facing, and mistakes there carry real consequences. The enjoyment filter protects the parts of the job people actually like, because automating something you love is just spending money to make work less interesting.
Her own example from the day this episode recorded: she built a script to pull LinkedIn post metrics (impressions, comments, likes) into Notion. Before that, an assistant handled it. Before that, she did it herself. She describes the task with open contempt, which makes it the perfect candidate: something done constantly, where imperfect output is acceptable, and which requires 0 joy to hand off. She calls it boring is sexy. “Figure out the workflow you do repeatedly, and then if mistakes are manageable and you’re okay with them, delegate and automate with AI.”
People get frustrated when they hear this. You show up to a bootcamp or hire a consultant expecting to leave with something impressive. Instead someone hands you a whiteboard. But Tata is direct about the tradeoff: “It takes time and it slows you down, sort of feels like it slows you down. In fact, it speeds you up.”
The same logic applies to how people first explore AI tools. Pure tinkering has value: testing a new model, playing with a capability outside any work context. That’s curiosity, and it’s worth protecting. But when something needs to work reliably in your actual job, setup is non-negotiable: context files, folder structure, clear instructions. The AI can’t fill in what you don’t give it.
The most durable AI workflows come from people who got honest about which parts of their week are boring, repetitive, and low-stakes. LinkedIn will give you inspiration. Your Miro board will give you your actual starting point.
Key takeaway: Map your actual workflow before opening any AI tool. For each repeated task, ask whether mistakes are acceptable and whether you actually enjoy doing it. Frequent, low-risk, low-joy work is the right first target. Build from there.
Back to the top ⬆️
Why Deep Generalists Outperform Channel Specialists in Marketing
There’s a running debate in marketing about whether to go deep in a specialty or build broad across domains. The specialist argument has genuine weight: if you’ve never actually run an SEO campaign, how do you know when an AI is confidently producing garbage? Tata sees the point. She also thinks the framing is wrong. Specialization built around channels is the vulnerability, and channels keep changing.
Her term for what marketers should actually become is “deep generalist,” a phrase she found on the internet and adopted because it captures something the T-shaped marketer framework mostly misses. A deep generalist has real expertise in at least 1 domain but deliberately builds breadth around it. The depth is still there. The difference is the deliberate horizontal stretch.
“The issue I’m seeing a lot of people are having is because they don’t know how to say if it’s a good or bad enough thing. So one of the things for us to train is judgment and taste.”
She watches this compression play out in her bootcamp every cohort. At the start of cohort 6, a participant said her team of 4 had been cut to just her. As the remaining content writer, she was now responsible for everything: SEO, social, website, the whole thing. That’s not a future prediction. It’s already the operational reality for a large share of the marketing workforce, and the people who trained deep in a single channel with no adjacent experience are the ones struggling most.
The channel argument is where Tata’s case gets sharper. An “SEO specialist” built around Google search has a real problem now that AI Overviews are reshaping how search works. Nobody building a “TikTok specialist” career a few years ago expected it to become a top-performing B2B SaaS ad channel. But 1 VP of business development recently told Tata that’s exactly what’s happening at their company. Channels are fluid. Betting deep on any specific 1 locks you into an increasingly narrow position.
Her own example: at Picsart, 1 division had no SEO function and no budget for an agency. Tata spent 2 months doing the SEO work herself, learning enough to direct AI through the process. When the business eventually hired an SEO agency, the agency was impressed by what was already in place. She had put in enough time to know what good SEO looked like and how to direct AI against that standard effectively.
The underlying skill that makes all of this work is judgment. Generating an image is table stakes. Knowing whether it’s good, whether it fits, whether an agent’s output is trustworthy enough to use: those require domain awareness that a specialist might have in their lane but that a generalist applies across the whole system. The specialist catches errors in their lane. The generalist sees what’s failing between lanes.
How to Build Genuine Competence Across Multiple Marketing Disciplines
The natural follow-up question is how much practice is actually enough. Tata’s answer is both specific and inconvenient. Formal credentials are largely irrelevant here, because the world changes faster than any curriculum can track. What she looks for is evidence of practice and real results.
“Even with AI, her research was really mediocre in the first couple of months. It took her about six months. And now she’s doing an amazing job. But it took her, even with AI, about six months to get there.”
Her rule of thumb is roughly 100 hours of real work in a domain. Her research assistant, a university student who had never done research before, is the illustration. Even with AI tools and structured workflows, the output was mediocre for the first few months. It improved steadily through feedback. After about 6 months, it reached genuine quality, good enough to function but not yet investor-grade. AI accelerated the process. The reps still had to happen.
For marketers who want to expand their range without knowing where to start, Tata’s concrete suggestion is to build something for yourself. Start a podcast. Write a newsletter. Post consistently on LinkedIn. Each requires managing production, audience, format, and distribution. Do them together and you’ve covered more of the generalist skill set than most marketing degrees touch, and you end up with a public portfolio that shows your range rather than just claims it.
Key takeaway: Identify 1 adjacent marketing domain you’ve been avoiding and spend 100 hours on real projects in it over the next 6 months. Take on tasks outside your formal scope, run experiments on a side project, or build something you own. Track the feedback carefully. Competence that transfers comes from doing the work, seeing the results, and adjusting based on feedback.
Back to the top ⬆️
The Rise of Diamond Org Charts
LinkedIn tells 1 story about flatter marketing orgs. Inside the companies Tata consults with, a different story is playing out.
The “organizations are getting flatter” narrative is true in a narrow slice: scale-ups with a founder who thinks in systems, who sees marketing and go-to-market as a workflow of inputs and outputs rather than a headcount problem. Tata has seen it happen. She recently exchanged notes with the founder of a financial scale-up called Entropy, who actually restructured her org and wrote publicly about the process. Those founders exist. They’re the exception.
“I don’t see flatter orgs that often. I see the job responsibilities of a person actually inflating and becoming sometimes ridiculous.”
Large enterprises are a different category entirely. At companies the size of Nike, Deloitte, or Nestlé, the structure isn’t changing much yet. What’s changing is the job posting. AI evangelists, AI task forces, and AI literacy requirements are appearing across every major company, even in traditionally conservative markets like Switzerland. The actual org shape is a different story. Tata was speaking at a conference when someone approached her afterward: “I work for a large organization. I want to learn, but they aren’t letting us in to build anything.” That’s centralized experimentation with everyone else watching.
The theoretical evolution most discussed is a shift from pyramid to diamond: fewer execution roles at the base, a larger coordination and judgment layer in the middle, management largely intact. The logic is that AI handles the execution work that once required a team of specialists. It’s a plausible model. In most places Tata works, it hasn’t arrived yet.
What she actually sees most often is founders looking at AI capabilities and concluding that their existing team should produce significantly more. The time savings from AI tools aren’t being used to reduce headcount. They’re being used to expand scope. The sales manager now also runs marketing. The marketing coordinator handles product launches, influencer programs, and analytics. Job descriptions grow to fill whatever space the tools create, and nobody stops to ask whether the organizational structure around all of this still makes sense.
Adding scope to the same headcount without changing the structure is deferred headcount planning. The marketers Tata sees struggling most are the ones whose job descriptions have quietly tripled while the org structure around them stayed identical.
Key takeaway: Audit your team’s role scope before assigning new AI-enabled work. List what each person owns today, identify what AI has freed up, and map what you plan to add. If the net result is a significantly larger job description with no structural change, you have an org design conversation to have. Use the AI capacity to do fewer things at higher quality.
Back to the top ⬆️
AI Tools Need Change Management
The easy assumption is that the main obstacle to AI adoption in marketing is leadership resistance. In Tata’s experience, the CMOs and VPs she works with are generally eager. The harder problem is the floor below the announcement.
She mostly works with CEOs and CMOs at scale-ups to identify where AI belongs in their workflows and how to run the rollout. One Swiss client stands out: a CMO who came to Tata because her team was resistant, and an external perspective seemed more likely to create movement. Tata spent several days with the team, looked at how they actually worked, ran the whiteboard exercise, and did something most AI consultants skip: she created a safe space for the team to say what they were afraid of. In that particular company, the jobs genuinely were not at risk. Saying so credibly changed the dynamic.
“I cannot stress enough how many times I walk into the room and on paper the company has Copilot access, but then there is only 1 person out of 10 people who got the license and they’re not allowed to use anything else.”
Once the fear was named and addressed, the tactical work started: tutorials, hands-on training, building actual workflows with the team in real time. People began shipping. That sequence matters. Start with the system change without addressing the fear and the tools stay unused. Tata is direct on this: most organizations don’t have enough AI-native marketers to replace a team even if they wanted to, and the existing team holds institutional knowledge and domain expertise that no new hire can replicate quickly.
The mechanical failure she sees most often is simpler than any of this. She regularly enters companies where the official answer is “yes, we have AI tools licensed” and the reality is that a small fraction of the team can actually access them. Software licenses are the floor. Before any change management strategy, before any training program, check how many people on your marketing team can actually open the tools you want them to use.
The model she points to as a smart leader move came from Olga Shydlovska, a former VP of marketing at SEMrush. When rolling out AI to her team, Olga asked everyone to list the tasks they found most tedious, then built the first wave of AI tooling entirely around those tasks. Skip the exciting use cases for now. Remove the friction from the boring work first. The team experienced relief rather than threat. That framing, Tata says, is more powerful than most leaders realize.
Key takeaway: Audit how many people on your marketing team have active licenses for the AI tools you expect them to use. If the answer is fewer than most of them, fix that before planning any training or rollout. Then identify 3 tasks your team finds genuinely tedious and build the first AI workflows around those. Fear of replacement drops when the first thing AI does is remove work people hated.
Back to the top ⬆️
How to Keep Your Marketing Job When AI Is Automating More of It
The question comes up more than people admit. A CEO of a company doing $100 million in annual revenue told Tata last week that he was afraid to automate his workflows because it might make him redundant. The morning of this recording, someone on a client call asked her directly: “Are you not afraid that you’re doing all of this AI work and you’ll no longer be needed?” The fear isn’t irrational. Someone Tata spoke to recently had a friend who got fired specifically because he automated so much of his own job that the role disappeared.
Her actual answer is less reassuring than most, and more honest. If the core of your role is structured, repeatable, and entirely deterministic (process A leads to B leads to C), then AI is genuinely capable of handling it. And if that’s most of what you do, upskilling isn’t optional. A harder question sits underneath the practical concern.
“If my job could be fully automated, why am I doing that? Like, generally, what’s the purpose?”
The second part of her answer is about visibility. The most technically talented marketers Tata has worked with are often the least visible inside their own organizations. They’re not building in public, and even within their companies they’re not communicating what they actually do or how their skills transfer to new problems. The value is there. Nobody is seeing it. She notices this pattern particularly among women: so many of the most capable people she works with don’t think they’re good enough, and they’re not making the case for themselves even internally.
The designer example makes the dynamic clear. When AI image and video tools started improving rapidly, designers in Tata’s network were among the most anxious. The concern was real. AI can prototype fast, generate variations at scale, and produce content that’s technically competent. The designers who ran into trouble were the ones whose value was primarily execution speed.
The ones who are thriving are the ones with taste and judgment: specific visual instincts, an eye for what works in a specific human context, a feel for UX flow that AI produces generically. “The bad designers who used to scribble really random stuff, maybe they’re out of business soon. But the ones who are good and who have taste and judgment, they have more work than anyone else.” Her designer contacts are booked through the end of summer.
Key takeaway: Write down the 3 things you do at work that no AI tool could replicate without you: your judgment calls, your client relationships, the institutional knowledge only you carry. Make sure that list is visible to your manager, your team, and the people who make decisions about your role. Visibility is survival, and the marketers struggling most are the talented ones nobody can see.
Back to the top ⬆️
Start Implementing AI Voice Diary
After an hour of talking about AI systems, org design, and automation strategy, the show closes with the question it always closes with: how do you decide what deserves your energy at any given moment, and how do you stay aligned with what actually makes you happy? Tata’s answer is less philosophical than you might expect, and more practical.
She uses a voice diary. A few times a week, she records short voice messages to herself about what happened and how it made her feel. On the weekend, she listens back. The written journal felt too analytical. The voice memo catches something the journal entry misses: the emotional register underneath the words.
“I have my voice diary and I send myself messages, and share what happened throughout the week and how that made me feel. I could hear my voice going a little sad or unhappy at some points. And that oftentimes tells me if I’m aligned with what I wanna do.”
The second signal is energy in real time. She pays attention to how she feels before and after different kinds of work. Preparing for this conversation energized her. Some client calls do not. When something feels off, she acts on it. She has walked away from potential clients whose vibe on an initial call felt wrong, even when it meant leaving revenue behind. She learned that lesson the hard way: ignoring the signal and taking the engagement anyway led to weeks of stress that cost more than the revenue was worth.
And then there is Sheba, the dog, who handles roughly 90% of the remaining problems. For anyone working from home full-time who has not yet acquired a pet: Tata recommends it without qualification.
Key takeaway: Record a short voice memo at the end of each workday for 2 weeks. Listen to them on the weekend. Pay attention to where your voice drops, tightens, or goes flat. The emotional data in your own voice is more accurate than any audit you’ll do in your head, and it will tell you faster where your energy is actually going.
Back to the top ⬆️
Episode Recap
The argument Tata makes is that good judgment in marketing is increasing in value faster than specialized execution. AI is handling more and more of the execution layer. What it can’t do is decide which workflows to touch, evaluate whether the output is actually good, or read the room when a client feels off. Those things require pattern recognition built from real work across multiple domains. That’s what a deep generalist develops that a channel specialist, staying narrow, increasingly doesn’t.
The tactical thread connecting the chapters is the 3-filter framework: frequency, acceptable error rate, and low joy. Tata applies this with every client and every bootcamp cohort. The best automation candidates are almost always the boring, repeatable tasks nobody argues for keeping. Interesting, high-stakes, or enjoyable work gets protected. Everything else is fair game, and the marketers who start there build momentum that the LinkedIn-inspired use cases never produce.
The bigger-picture implication is about org design, and it’s uncomfortable. Most companies are not getting flatter. They are expanding individual job scope and calling it efficiency. The marketing manager who now also runs influencer programs, product launches, and analytics is not working in a flat org. She is working in an underfunded organization that found a way to defer the structural conversation. Tata sees this across scale-ups and large enterprises alike, and she is direct that it is a problem compounding quietly across a lot of teams right now.
The honest admission running underneath the episode is about fear. The CEO of a $100 million company asked Tata last week if automating his workflows would put him out of a job. It’s a real question, and she takes it seriously. Her answer is that if a role can be fully automated, it was always a fragile foundation. The more useful move is to identify what cannot be automated (judgment, taste, institutional knowledge, client trust) and make sure those things are visible to the people who need to know about them. The most talented marketers she knows are often the least visible inside their own organizations. That’s the gap worth closing.
Connect with Tata on LinkedIn at linkedin.com/in/tatamaytesyan or learn more about her AI in Marketing Bootcamp at Maven.
Full episode ⬇️ or Back to the top ⬆️
✌️
—
Intro music by Wowa via Unminus
Cover art created with Midjourney (check out how)
Event Shoutout: Demand & Expand 2026
Join the world’s best B2B marketers for 2 full days of tactical content, in-depth sessions, and meaningful connections. Use HOM20 at checkout
Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS
Related tags
<< Previous episode
Next episode >>
All categories
- AI (104)
- career (65)
- customer data (64)
- email (64)
- guest episode (177)
- operations (127)
- people skills (34)
- productivity (10)
- seo (15)
See all episodes
Future-proofing the humans behind the tech
Apple •Pocket Casts•Google •Overcast •Spotify •Breaker •Castro •RSS