left arrow

How to survive as a SaaS in an AI-first world

April 17, 2024

Des Traynor, Chief Strategy Officer at Intercom, came to Hexa a couple of weeks ago to talk about the urgent need to account for AI in software products. Here’s a summary of the main points covered.

According to Intercom's Des Traynor, in 6 to 9 months, everyone will be using AI, and it will be essential for every software company that wants to stay alive to integrate it. You’re either an AI company or you lose your entire business to someone who has built an AI version of your product. You have two jobs: survive in AI, then thrive in AI. AI will change so much in ways we do not yet even comprehend. Therefore, surviving is a challenge in itself, one that many will surely fail.

Survival depends not on being the biggest or the strongest, but on adapting quickly.

AI will disrupt ALL software products

If you think AI won’t impact your business, you're preparing for failure. There is very very little that AI can’t massively simplify, or cut out altogether. Consider an ad optimization platform: today, it requires logging in, creating ads, manually reviewing them, selecting the best, and optimizing based on performance. AI can whittle down this workflow to a simple approval on the user’s end.

AI could even transform how we use tools like Excel. You will no longer need to manually input formulas or do time-consuming tasks such as applying conditional formatting. Instead, you can simply tell AI your requirements—like “highlight all negative numbers in red”—and it will handle these tasks instantly. This not only speeds up work but also reduces errors, making processes smoother and more reliable.

… but some more than others.

There is a spectrum of how much AI will change software—some product categories will be completely transformed or even become obsolete, while for others AI will be a way to enhance current features without big changes to the core of the product offer. At the former end of the spectrum, startups will win. At the latter, incumbents will win. In the middle, it will be a race.

A morning with Des Traynor at La Cristallerie

Startup opportunities lie in the ‘disruptable’ areas

Startups should focus on the ‘disruptable’ end of the spectrum, where AI will trigger a complete overhaul of what already exists, where entire tech stacks will need to be remade, and workflows completely rethought. If you try to innovate at the other end, there is a high chance your work will end up becoming unpaid R&D for incumbents.

Even if a startup has a significant headstart, as soon as the incumbent in the area wakes up, it is likely they will take over. For example, developing an AI-driven email platform might seem like a good idea, but incumbents like Mailchimp could swiftly overpower new entrants once they decide to integrate AI into their services. The key challenge is to identify niches that are ripe for disruption, where startups can establish themselves as leaders.

… not in products replaceable by Apple’s next OS.

Operating systems are another competitor startups need to account for. There could be a future in which apps get consumed by AI, subsumed into a single LLM where you ask Siri to do everything for you. Instead of going via Uber Eats to order a pizza, for example, you may just ask Siri. So, make sure your startup’s product can’t be replaced by this.

To sum up, if you’re a startup thinking about AI, focus on areas where:

  • The technology stack is likely to be extensively revamped due to advancements in AI, minimizing the advantage of existing players.
  • The benefits you offer aren't readily available 'off the shelf' from companies like OpenAI or Anthropic.
  • It is unlikely that operating system providers will automatically include a strong competitor.

One day ⏩ day one

If there’s one lesson you need to take away from this, it’s that companies need to wake up to AI right now. For every software business, young or old, this isn’t a distant future. You either lead by integrating AI today, or you will inevitably fall behind those who do. The choice is stark: adapt or die.

If you want to help build the future of AI startups with us, find out more.