The execs guide to blowing a billion on AI

“A million dollars isn’t cool. You know what’s cool? A billion dollars.”

The Social Network (the movie, not the concept) is 15 years old. The quote is Napster CEO Sean Parker telling Facebook founder Mark Zuckerberg to think bigger. Well, that feels pretty small now. The latest fund-raise from OpenAI was $40bn, just to keep development going; the plan is to spend $500bn over the next few years to build out AI infrastructure.

“$7bn in training costs in 2024 alone” cool? “Total losses from 2023-2028 projected at $44bn” cool?

Well, maybe: if this revolution pays off, the possibilities are limitless. Big bet, big return.

But for almost all businesses, even throwing much smaller sums of money at AI projects is a heck of a ‘maybe’. The big consultancies will say it’s the table stakes for competing into the 2030s. Another ‘maybe’. (Even Boston Consulting Group says only 4% of AI experiments are generating real value from their budget.)

Pretty quickly we can see that building out your own AI architecture is both risky and bone-crushingly expensive.

Mattresses, not moon-shots

But there are things you can do – right now – that exploit the investment that others have made in AI. Taking small parts of the AI revolution – the bits we understand and can control – and applying it as part of your workforce solution can work, and without making huge financial and operational bets.

Exploiting advanced, expensive research for profit has a pedigree. For example, NASA’s Ames Research Center developed what is now called ‘memory foam’ for better crash protection. Now it’s used in homes all over the world for comfort, and thousands of businesses profit.

Our version of that for AI is the multi agent workforce. Instead of trying to reinvent a business for the ‘AI age’ (which, OK, is profitable if you’re a huge IT consultancy working with outsized S&P500 budgets) we work with companies of all sizes to analyze where AI agents can work in high-potential processes alongside humans to deliver efficiency and agility.

Two things drive our approach.

First, we think technology that enables people to work smarter should deliver ROI pretty much straight away. Sam Altman at OpenAI can lose $44bn in pursuit of artificial general intelligence. Most business want to drop smarter tech into their workflows and immediately free up people’s time and do better work.

Second, we know that AI feels risky to a lot of people. And they’re right to be nervous: throwing AI at processes doesn’t work, and is a governance nightmare. That’s why our approach is modular, audited, involves testing, is transparent… and evolves over time.

AI development is moving fast – and the beauty of the multi-agent workforce is that it is designed to evolve with it, refining tasks and processes with human inputs. The AI gets smarter? You refine a single task or the agent that completes it. Your business changes? You’re not retooling an AI IT project from scratch – you’re looking at what agent (human or AI) can meet your new needs.

A billion dollars is less cool these days than it used to be. We think ‘cool’ for most organizations is deploying reliable tech, in the service of better outcomes for your people and your clients, that boosts your profits and makes you more competitive… fast.