Your company's LLM token spend is out of control, and getting more expensive by the day. What do you do?

At the onset of the "tokenmaxxing" wave, several people waved off my concerns of cost overruns. They told me that it was perfectly OK to throw money into LLMs without any kind of checks in place.

Their assumption? That token prices would fall, so they would only see a short bump in their spend and it would eventually drop.

Fast-forward to today and companies are having second thoughts on that approach. Those monthly LLM bills are growing out of control and there's no relief in sight. Some companies are even thinking of cutting developer headcount to make the balance sheet line up.

So what do you do?

Your first move might be to cap everyone's LLM token spend. That's a good idea, but that's your second step.

Your first step is to get a handle on all of your genAI projects. Figure out which ones show promise and which ones are just money pits. It's time to close the money pits.

Once you've narrowed your project focus, then sort out token caps and other cost-trimming measures.

Point being: capping usage across the board will starve worthwhile projects while letting wasteful projects continue. You'll save a little money but your larger genAI plans will suffer.

▶️▶️ Want to get your company's AI efforts under control and on-track? Reach out. I can help.