The new era of AI business models
"This time is different"... but for real
AI tools are beginning to proliferate in our every day lives. While most of these tools are reaching us via Big Tech trying to upsell existing product lines (e.g., Microsoft Copilot), startups are beginning to tinker with new business models that could disrupt many incumbent business today.
So far, two main business models have emerged that AI companies are leveraging (both of which we discussed last year):
Usage Based Pricing (largely used by Foundation Models): Charge customers to access your best models on a per token or per API-call basis
Seat Based Pricing (largely used by AI Copilots): Charge end users on a per seat basis to enhance their productivity at specific tasks (not dissimilar from vertical SaaS tools)
While these models are working for many startups, they don’t always make sense for many others. As AI helps to reimagine the way that we do work and interact with others in the business world, many savvy entrepreneurs are revisiting the core “Jobs To Be Done”, and experimenting with new business models in the AI era.
Some of these are in the very early innings of experimentation, but I believe will be transformative to the global economy. A few of the business models that I’m most excited about include:
AI-as-a-Service / Outsourced Services: Many AI tools have begun to emerge that make white collar services workers (e.g., Lawyers, Accountants) more efficient. But why let them capture all the value? Why shouldn’t there be a new era of AI-enabled services firms that emerge and disrupt analog services firms themselves?
In this case, I’d expect customers to pay on a per project or per deal basis, similar to how they’re already accustomed to interacting
Employee in a box: A few weeks ago, Devin AI took the world by storm by positioning itself as the world’s first autonomous software engineer. As agents emerge in other fields, will we work alongside AI’s similar to how we interact with our human teammates?
This has some businesses beginning to think of AI tools as an alternative to human heads. Does this mean their annual “salary” begins to match that of local talent? (How will big corporations retain their AI employees if they can’t offer them pizza parties in place of annual raises…?)
Commission-based businesses: In a few new businesses, I’m beginning to see founders overcome customer skepticism by taking the risk on success. Would you be willing to adopt an AI procurement tool that only gets paid when it saves you money? Chances are, yes!
For startups that are confident in guaranteeing cost savings or increased sales, why not try a commission based pricing model? This allows startups to share in the upside that they create
So what’s the catch?
All of these are fun thought exercises, and I am far from the first person to highlight some new potential business models. But the world doesn’t change over night. As exciting as these business models are, they don’t come without strings. While all come with the promise of faster customer adoption (and therefore topline growth), the opposite is also true.
As we have discussed in the past, Tech companies are valued on ARR. Primarily, the reason for this is that it’s highly recurring, high margin, and easy to forecast, so you can sort of think about it as an annuity on top of your fixed cost base.
Enterprise SaaS companies typically sign multi-year contracts with their customers at low churn rates. For these new business models, however, you lose the ability to guarantee your revenue forecast.
Just like any law firm, individual employee, or procurement agency, your compensation isn’t guaranteed every year. There’s lumpiness, seasonality, and economic cycle risk embedded in each of these markets.
As such, you historically have not been able to value any of these new businesses models on an ARR basis.
This has the dual impact of driving both VC’s and Founders crazy, as nobody really know how to value these businesses. Startups typically need to raise capital to function, and don’t generate earnings for a long time.
Even if the growth is great, how do you justify investing in an unprofitable business that can’t be valued on ARR?
We’re still in the early innings of figuring this out, but I suspect it means that these businesses will have to raise at more modest valuations to begin with, and will have to prove out unit economics and profitability earlier than their Enterprise SaaS peers.
For high margin services businesses that prove that they have great customer retention, maybe investors can get comfortable with ARR multiples as they scale (or at least, assigning premium EBITDA multiples).
However, the size of the prize in these spaces is huge. The global services market dwarfs that of the enterprise software market. If any of these businesses really get on a roll, they have a possibility to become deeply defensible global behemoths in long run.