AI is one of the fastest-growing startup sectors globally. In May alone, large language models raked in $12.4 billion, according to Crunchbase. To put this in perspective, this represents 40% of the global total raised in May.
But a disquieting truth lurks beneath the surface. While the initial hype was deafening, there are signs it’s fading. The question now hangs heavy: can high-flying AI startups translate early promise into long-term success?
As a venture capitalist, AI-native startups are easily some of the fastest-growing startups I’ve encountered, scaling to $1 million in annual recurring revenue in record time. According to Insight Partners, nearly half (44%) of new unicorns in 2023 were generative AI companies. The early-stage growth is undeniable, but the meteoric rise often masks a crucial challenge. In this case, it is the lack of product planning past the demo. Demo value doesn’t equal long-term business value.
The recent excitement around AI bears an uncanny resemblance to the “Peak of Inflated Expectations” stage in Gartner’s Hype Cycle, where the emerging technology’s early publicity has produced several success stories accompanied by many failures.
With AI hype starting to come back down to earth, we are on the precipice of the next stage — the “Trough of Disillusionment”. This is where overinflated expectations both from the customer and investor side are met with reality. Customer and investor mindshare slows. AI startups will be under more pressure than ever to show results and revenue.
Now is the time for founders to deliver against what they said they would and match against customer expectations. From there, they can continue to ship products to ensure that the company doesn’t plateau or, worse, crater. To land successfully, AI startups must consider three key stages: the initial launch, flight mode, and the descent into the world of growth and expansion. Own more of the workflow and data to drive to a system of intelligence.
Compared to traditional startups, there are lower development time barriers to entry for AI application layer companies. AI startups require less initial capital investment due to the availability of foundational models through an API and open-source models. AI’s potential for disruption and high returns has also attracted significant investment from VC. This means more customer value in a quicker timeline, but also more competitors vying for your customer’s attention. So, founders need to think beyond the initial wedge.
There are four key questions that AI founders need to ask themselves to show they can think beyond the liftoff point and sustainably grow their businesses:
Working backward, what is the most ambitious version of the company, and what experiments can we run to maximise our learning to get there in the next year?
Why is this the right wedge product/market, and how do we build from there? Can we build to cater to more of the workflow and help deliver more value to customers over time?
What logical steps can be proven out or derisked over time to show that you’re on the right track?
What unfair advantage does this give you over time? Why would it be hard for a customer to leave you?
If a founder can answer the questions above, they can build a more durable business over time and accurately track and measure the resources required to reach the next stage. Quick successes are great for a dopamine hit, like a sugar rush for your business, but without a roadmap for what comes next, all the initial effort fizzles out.
Now, the real test begins. To achieve sustained orbit, founders must shift their focus. Capital is important for lift-off, but on its own, it very rarely leads to success. Capital is a tool, not a strategy. Success is achieved by properly executing a strategic vision.
First, ready your crew. Founders must assemble the right team and talent to navigate the complexities of AI development and business growth. AI and machine learning talent are scarce, but great developers can bridge the gap between regular development and work with LLMs. Hiring great system engineers and architects can save you a lot of headaches and help you scale. Since team construction can be small, having people who understand the technology but have an appreciation for the business side and customer value will help you build the right things and move quickly. While excited, many industries have been stuck on the start line and haven’t fully explored how AI can improve their operations or workflows. This opens doors for startups with niche solutions tailored to specific sectors. When you engage with customers, you know them so well you can read their minds and know their needs. That depth of understanding goes a long way.
Second, refine your product. Remember, that initial launch was just the first iteration. Now, it’s time to gather data and user feedback to optimise your offering. Focus on demonstrating a clear return on investment for your target audience. This is what will keep them engaged and coming back for more.
The third stage is a descent into a world of sustainable growth and continued expansion.
Staying on top of your cost structure to new revenue. You can use the ‘magic number’, which is the net new ARR in a period divided by S&M expense from the prior period. Ideally, the ratio is greater than one.
Next is to identify growth opportunities. Can you scale your solution to new geographical markets or user segments? Are there strategic acquisitions that could complement your core offering and accelerate expansion? A multi-stage plan shouldn’t be static. It should be a dynamic compass that adjusts to the ever-evolving landscape.
This involves understanding the repeatability of your existing business and foresight about industry trends, such as technological advancements or shifting consumer preferences. Take stock of your existing assets and resources — the tools and talent that make your startup unique — and develop a keen eye for adjacent opportunities that can add value and propel further growth. One exercise to run is what Airbnb used from Snow White’s example to plot all user journey interactions that lead and move beyond your product today. You will be surprised you can start building a more expansive experience to help your customers.
When refuelling, remember that fundraising at this stage isn’t about plugging holes; it’s about propelling your fuel into an efficient flywheel where you have reasonable predictability. Putting fuel into a rocket going in the wrong direction, unfortunately, only accelerates you away from where you want to go. Focus on showcasing metrics that quantify your success, not just the initial hype. It is important to align with the right investors and partners who can provide the strategic direction and networks to empower founders to navigate the path to success.
Much like rockets, AI startups are experiencing exhilarating liftoff and have huge potential for stellar heights, but landing requires precision and a clear roadmap.
Never miss a story: sign up to SmartCompany’s free daily newsletter and find our best stories on LinkedIn.
Andrew Poesaste is a partner at venture capital firm Rampersand.
