AI Sprawl is Becoming a Problem

“Which AI tools are you using these days?”

This has become the new icebreaker. I’ve learned to limit it to one or two because people don’t want to hear the full list. They’re looking for validation of the tools they’re using, and maybe some suggestions of emerging tech that they may have missed.

The AI Tools list has quickly become as complex as Scott Brinker’s Marketing Technology Landscape Supergraphic. The 2025 MarTech list includes over 15,000 tools.

Unfortunately, due to the speed at which AI tools are being released, it’s impossible to measure how many AI tools there (if only there were an AI that could keep track of that). The best I’ve seen so far is the AI Industry Mindmap that has 50,000 companies listed.

Best-of-Breed vs. All-in-One

When I’ve overseen marketing technology, the question always came down to best-of-breed or all-in-one tools. There are pros and cons to both. Best-of-breed means you get the best tool available, even if it natively doesn’t work well with other tools in your stack. The all-in-one tools will (usually) work together, but most of the tech stack ends up being far from the best.

The other struggle with best-of-breed is that you end up with functionality overlap. At one point, I oversaw a MarTech stack that included 60-some products. There were similar capabilities of different levels across multiple tools. Most of these tools do one thing really well, and a few things are only so-so, and it’s because an important customer said they wanted it built into the tool. You end up paying more, using less, and watching your stack bloat into Frankenstein’s software monster.

AI Tech Stacks are Getting Just as Bloated as MarTech Stacks

There are specialized companies out there that do amazing things. These companies have to fight two battles: staying ahead of similar competition and becoming a feature in one of the bigger AI systems. Meeting transcription is a perfect example.

Multiple companies use AI to record meetings and allow you to interact with the conversations to ask question, get a list of action items, and even determine the tone of the conversation. I was an Otter subscriber. Granola is another one that came to me highly recommended.

ChatGPT’s new Record feature is good, but not great. I couldn’t get a full transcript from a recent meeting, which means I couldn’t feed it to a custom GPT. Otter and Granola are safe … for now.

As ChatGPT rolls out more features, like the recently launched ChatGPT Agent (I just got access on Thursday!), it is becoming more of an all-in-one tool. And it’s eating more of the AI market. There’s a good reason for that.

It’s All About Branding

Here’s why that matters: in tech, function matters, but brand dominance decides who survives. For those who don’t use these tools on a regular basis, ChatGPT is the brand for AI. My mother uses ChatGPT. When people are starting to explore AI, they’re going to go with the name they’ve heard the most. These folks don’t have to worry about AI sprawl because they’re only using one or two tools at most.

I was in a meeting earlier this week, and I was asked if I ever use Apple’s AI. I laughed because it’s still borderline useless. I know there is a lot of AI happening behind the scenes on my iPhone, but as far as using Apple Intelligence, it’s unlikely. Apple has crushed its brand in the AI space because it made a big deal out of AI, and then didn’t deliver (which led to class-action lawsuits).

What’s in Your AI Tech Stack?

So, how many tools are you currently using regularly? This question is tricky because you are probably using tools with AI embedded into them, like Gmail or Adobe Reader. But for AI-specific tools, I’m using a few like ChatGPT, GitHub Copilot, Midjourney, and others. How about you?

Booking for August

I have space for one more consulting client. If you need help with digital strategy, creative direction, AI and marketing, or brand development, let’s talk.

And if you’re building out a team and looking for someone full-time, I’m open to the right opportunity.