Buyer , User , Startups — 07.11.2023

Use AI to Build Something Better — Not Something New

At a Glance:

  • Founders are receiving confusing investor advice about how to implement AI. As a result, they’ve been struck by decision paralysis and don’t know how to move forward.
  • Founders shouldn’t ask, “what new thing can AI help us do?” Instead, they should ask, “how can AI help us do what we’re already doing better?”
  • Investors can help founders by connecting them with AI industry experts who’ve been building AI products for decades (yes, they exist).

AI feels new to founders, but it’s not different. 

Recently, founders have been asking me the same question: “how do I incorporate AI into my roadmap?” They walk me through a variety of investor advice about how AI should be a part of their product. Advice ranges from hiring data scientists earlier to leveraging public LLMs like ChatGPT. 

I have to remind founders (kindly) that investors are mostly new to this technology, so investor advice is no less a guess than the founders.

Story time! I have a long history with technology. It fascinates me. In high school, I downloaded my first illegal copy of Photoshop. During my undergrad, I studied the history of technology and today, I continue to read histories of pixels, VC and technology. While I used to love learning about tech because I was a curious sponge soaking up everything I could, now it’s proving immensely helpful as I advise founders on navigating massive technological shifts like AI. 

Yes, AI has rapidly accelerated in the past year. With that acceleration, it can be challenging to see the river for the rapids. As the current increases, it’s easy to panic and forget the course you’re on. By looking at the history of innovation in technology, it helps me provide a better framework for how to handle any major tech shift.

And if you’re thinking, “but AI is different! It’s going to change the world,” you’re not the only one. While the tool itself is unique, the concept of a huge technological shift that causes decision paralysis and massive downstream effects is not new. A decade ago, it was cloud computing, five years ago it was blockchain and right now, it’s AI.

Here’s what founders can do about AI. 

To navigate the current evolution, I find two core principles helpful — surround yourself with experts, and use technology to enhance the consumer experience, not just the product.

  1. Surround yourself with experts. Founders must continue focusing on their roadmap and product development. I’ve seen many of them become paralyzed because they’re spinning their wheels figuring out what AI means for their business. As they dive deeper into this technology, more time is wasted — they may “wake up” a month later with more knowledge of AI and LLM technologies but their product will remain unchanged. While AI is new to many founders, it’s not new in the tech industry. For decades, companies have implemented AI and universities have offered data science. AI is having a moment, for sure, but to many engineers it’s not new. Founders should seek out advice from seasoned pros or product experts (like myself) rather than investors. This type of advice can save a founder months of research, development and bad decisions. 
  • Think of how the technology can elevate your consumer experience, not what the technology can do for your product. My mentor in design school, Marty Siegel, called this concept, “computer imagination.” He taught designers to utilize the magic of technology to provide better experiences. Technology has innate properties that lend themselves to be utilized in different ways. With AI, ML and LLMs in general, the magic is more real than ever before. I remember the first time I saw a CRT monitor paint a color image of a cheetah and felt like there was a wizard in my school library. That’s the same feeling that we’re collectively having with the amazing ChatGPT prompts today.

Here’s what it means for teams.

While I am confident that AI will improve product team operations, I’m unsure how this technology affects team building. Every case is different and it’s not always clear when founders should hire their first data scientist. What I do know is that there simply isn’t enough available data science talent to serve the market. As the field gets more competitive, larger companies will scoop up the best talent, further draining the talent pool for early-stage startups. In reality, most startups probably can’t hire a data scientist FTE early on. As a result, founders should consider this and seek guidance from other sources like agencies, fractional product experts or even universities. 

But what about products?

For founders, the capabilities of today’s LLMs opens an overwhelming world of possibilities. Instead of imagining how AI changes the product’s value, think about how the product’s value can be delivered more effectively. 

It’s not, “what new thing can AI help us do?” it’s “how can AI help us do what we’re already doing better?”

When I talk to founders, this eases a lot of anxiety because it means they’re still on the right track and don’t need to shift their value proposition. 

When I conducted user research as a designer, one of my favorite questions to ask was, “if this product were magic, what would it do?” This gets the person to break the constraints of what was possible and to be selfish about what they really needed. 

Today, AI can (almost) deliver any answer to the magic wand question. Let’s look at a more specific example of this. 

In many B2B products, a dashboard or some sort of data interface exists to help users make decisions or to help management see the value they’re getting. Often this is done by delivering graphs upon graphs, assuming every user wants to look at charts. But when you ask them what they’d prefer, they’ll typically say something like, “I just wish the product could tell me what the data means without having to dig for it.” This is how AI is fundamentally prepared to help. It changes the delivery of the product’s value by doing exactly what the user wants. It can react to natural prompts, learn and deliver smarter results over time, and requires less and less intervention over time. 

AI shouldn’t change the value your product delivers, it simply helps you deliver that value quicker, better and more efficiently. 

Here’s what investors can do to help.

As an investor, if you’re able to counsel founders, tell them: 

  • While the tool itself is unique, the concept of a huge technological shift is not. A decade ago, it was cloud computing, five years ago it was blockchain and now, it’s AI. 
  • While AI is having a moment, it’s not new. There are industry experts who’ve been using AI and data science for years that could better advise you about your next move (especially related to team building).
  • The value your product delivers shouldn’t change. Don’t ask, “what new thing can AI help us do?” Instead ask, “how can AI help us do what we’re already doing better?”

Yes, AI will change everything in SaaS. But contrary to popular belief, the results won’t feel quite as dramatic as we think simply because our business needs are the same. Instead, AI is going to enable more magical experiences, and when used correctly, allow products to engender stronger engagement, require less usage, and provide more creative possibilities that we simply can’t see today. 

To hear more of my hot takes, subscribe to my newsletter.