1
0 Comments

Towards Result-Based Pricing in AI

Humans started off working with just their bare hands during the Stone Age. Eventually, we invented tools like hammers and wheels to ramp up efficiency and productivity. Innovations like the steam engine and the internet later caused huge leaps in productivity.

Fast forward to the 1980s, when humanity shifted from manual labor to setups enhanced by software. This meant humans still ran the show but used software to boost efficiency. While software traditionally just sped up tasks, AI has evolved to take over more complex parts of the job, much like coaching an intern. Yet, humans still need to oversee and integrate these AI-driven components.

Enter the era of specialized AIs like ChatGPT, leading us to the concept of autonomous agents—AI systems that collaborate to tackle intricate tasks without human help. As these agents become more capable, the business focus will shift from selling software to selling completed tasks. It also means companies will need to rethink their sales strategies. Instead of targeting IT departments, they'll be vying for budgets traditionally allocated to human resources. Soon, the game will be about delivering work at the best cost, quality, and speed—regardless of whether it's done by humans, with AI assistance, or fully autonomously.

And here comes my main point...

As the business model changes as we discussed, I anticipate pricing strategies will evolve too. We've seen shifts from fixed fees to flexible SaaS models, and from there to usage-based pricing. I think it's time AI solutions switch from focusing on input (sure, companies still need to cover their compute costs) to emphasizing output. So, I'm betting we'll soon see a move from pricing based on usage to pricing based on results.

A perfect example is Intercom’s AI customer support chatbot, Fin. You only pay when Fin actually resolves a customer’s issue. Specifically, you pay $0.99 for each "resolution", ensuring that the customer gets what he pays for and pays for what he gets.

Of course, this will likely be a gradual transition, but one I look forward to.

What are your thoughts?

on May 9, 2024