Outcomemaxxing > Tokenmaxxing

Something fascinating has happened in enterprise AI in the last few months.

Companies rewarded 𝘁𝗼𝗸𝗲𝗻𝗺𝗮𝘅𝘅𝗶𝗻𝗴 (use AI as much as possible), and now they're cancelling licenses as costs escalated.

That's the easy version of the story, and it's 𝗶𝗻𝗰𝗼𝗺𝗽𝗹𝗲𝘁𝗲.

This reminds me of a story from the 1800s. Officials hoping to cut the snake population offered a bounty for every dead cobra. But this led to an increase in cobra numbers. Why? Because people started breeding cobras for the bounty.

Then the program was scrapped. Sure enough, the breeders freed their now-worthless snakes, and the population ended up HIGHER than before.

𝗧𝗵𝗶𝘀 𝗶𝘀 𝗰𝗮𝗹𝗹𝗲𝗱 𝘁𝗵𝗲 "𝗰𝗼𝗯𝗿𝗮 𝗲𝗳𝗳𝗲𝗰𝘁."

Pay for a proxy, and people optimize the proxy, not the thing you actually wanted.

Is this what just happened with AI?

𝗦𝘁𝗮𝗿𝘁 𝘄𝗶𝘁𝗵 𝗪𝗛𝗬 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗽𝘂𝘀𝗵𝗲𝗱 𝘂𝘀𝗮𝗴𝗲.

Capital spending on AI infrastructure is staggering. When you've bet the balance sheet on capacity, the worst outcome is nobody uses it. So the rational move was to push adoption hard: usage targets, leaderboards, even ranking employees by tokens used.

Employees began running AI agents on trivial tasks just to inflate token counts. 𝗧𝗵𝗲 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲 𝗴𝗼𝘁 𝗮 𝗻𝗮𝗺𝗲: 𝘁𝗼𝗸𝗲𝗻𝗺𝗮𝘅𝘅𝗶𝗻𝗴.

Measure activity and you get activity NOT results.

Then the bills came in.

𝗕𝘂𝘁 𝗵𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗰𝗼𝗺𝗽𝗹𝗶𝗰𝗮𝘁𝗲𝘀 𝘁𝗵𝗲 𝘁𝗶𝗱𝘆 "𝗰𝗼𝘀𝘁𝘀 𝗸𝗶𝗹𝗹𝗲𝗱 𝗶𝘁" 𝘀𝘁𝗼𝗿𝘆.

I’ve seen Claude Code become the default choice for many engineers. They often prefer it to their own enterprise coding tools. So they use it more, and usage-based billing means costs climb.

Meanwhile enterprises need to show adoption on their own tools. The result is they cancel external licences.

Therefore, you can't call this a "cost only" story.

AND as tokens keep getting cheaper, the problem accelerates. That’s 𝗝𝗲𝘃𝗼𝗻𝘀 𝗽𝗮𝗿𝗮𝗱𝗼𝘅 in action: per-token prices fall while total bills climb, because consumption outruns price.

So the metric has to move from how much AI you use to what you get for it.

𝗖𝗮𝗹𝗹 𝗶𝘁 𝗼𝘂𝘁𝗰𝗼𝗺𝗲𝗺𝗮𝘅𝘅𝗶𝗻𝗴.

You can't fake an outcome by running busywork through an agent overnight.

𝗧𝗵𝗲 𝘃𝗮𝗹𝘂𝗮𝗯𝗹𝗲 𝘀𝗸𝗶𝗹𝗹 𝗶𝘀 𝘁𝗵𝗲 𝘂𝗻𝗴𝗹𝗮𝗺𝗼𝗿𝗼𝘂𝘀 𝗼𝗻𝗲: 𝗸𝗻𝗼𝘄 𝘄𝗵𝗶𝗰𝗵 𝘁𝗮𝘀𝗸 𝗶𝘀 𝘄𝗼𝗿𝘁𝗵 𝘁𝗵𝗲 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲 𝗺𝗼𝗱𝗲𝗹, kill a bad run before it burns the budget, structure the work BEFORE you spend.

This was invisible when tokens felt free, BUT it’s the whole game when they’re metered.

And let's be clear about what this is NOT.

It is not evidence that AI doesn't work. The tools are extraordinary and getting better by the month. The question was never the tech. It's what we measure and what we reward.

Fix the incentives, and this becomes the most powerful leverage any team has ever had.

Until next time,

Ram — 

Ram Srinivasan

MIT Alum | Author, The Conscious Machine | Global Future of Work and AI Adoption Leader published in Business Insider, Fortune, Harvard Business Review, MIT Executive Viewpoints and more.

—

A Message From Ram:

My mission is to illuminate the path toward humanity's exponential future. If you're a leader, innovator, or changemaker passionate about leveraging breakthrough technologies to create unprecedented positive impact, you're in the right place. If you know others who share this vision, please share these insights. Together, we can accelerate the trajectory of human progress.

Disclaimer:

Ram Srinivasan currently serves as an Innovation Strategist and Transformation Leader, authoring groundbreaking works including "The Conscious Machine" and the upcoming "The Exponential Human."

All views expressed on "Substrate" and across all digital channels and social media platforms are strictly personal opinions and do not represent the official positions of any organizations or entities I am affiliated with, past or present. The content shared is for informational and inspirational purposes only. These perspectives are my own and should not be construed as professional, legal, financial, technical, or strategic advice. Any decisions made based on this information are solely the responsibility of the reader.

While I strive to ensure accuracy and timeliness in all communications, the rapid pace of technological change means that some information may become outdated. I encourage readers to conduct their own due diligence and seek appropriate professional advice for their specific circumstances.