Campfire CEO/CFO on why the company doesn’t cap token use

On Oct. 22 at Campfire’s Basecamp summit in San Francisco, the company’s CEO and CFO John Glasgow walked on stage in a pair of custom Nike Jordan’s bearing the company logo.

Smiling, he told CFO Brew off stage that he paid an Etsy artist to paint them, and wears them to conferences. Glasgow and Campfire customers are planning for 2027 and thinking a lot about AI token spend and accountability.

“Typically there was always [a] software budget, and there was [a] payment budget. And where does token spend sit? And how do we think about usage-based pricing? Because if your largest bill is usage-based, unlike cloud vendors like AWS, it’s actually quite hard to forecast token consumption for an organization,” Glasgow said.

“But the big debate is do we cap token use per employee? What if the incremental tokens make them more productive? But we’re capping the use, and so actually, are we capping productivity?”

Managing the surge. An Accenture survey of 750 senior global executives between June and July found that their enterprises spent about $2.5 billion on AI tokens last year, and they expect AI spend to grow 78% over the 24 months from September. In addition, a little more than one-third (35%) have a handle on “cost per business outcome…for their single largest AI use case,” the Accenture survey found.

Finding a happy middle for token use—“we don’t want tokenmaxing, but we don’t want tokenminning,” Glasgow said—is a good way to decide what token limitations or budgets should be set at. But almost in contradiction, Glasgow said everyone at Campfire has unlimited token use.

“But just like using your corporate card, I ask everybody to treat tokens or use tokens responsibly—and if you’re [an] outsized token consumer on your team, just like corporate card spend, we’re really going to look at, why are [you] using more tokens than the rest?” Glasgow explained.

“There could be a good reason, if you are delivering outsized output…that’s how I ask employees in Campfire to treat tokens: It’s corporate funds, so that’s why we have an unlimited policy today, but it does require a high level of trust.”

Responsibility for output. One of the most challenging aspects of trying to regulate internal AI token use is accountability—or a lack of it, Glasgow said. He understands it’s in the best interest of the companies offering tokens to encourage token “munching,” or chomping away at large amounts of tokens quickly. But Glasgow said accountability starts with empowering the AI users themselves to be leaders of their own AI bots or models.

“I think the explosion of content and slop grenades, slop—there’s many phrases for it…How can humans continue to be accountable for the work that they produce? And if we’re all managers now, then your agent is on your team. I think even if you have an autonomous agent on your team, you are still responsible for its output,” he said.

“One of our core values of Campfire is actually ‘Quality with velocity,’ and I think the quality with velocity for leaders, now that we’re all managers, I think is becoming more and more important,” Glasgow pointed out.

With AI output, “because some employees candidly just get lazy, and they’re just one-shooting AI without reviewing it. And I’ll push back and I’ll say like, ‘Hey, did you actually read all this? It does not seem like thoughtful work,’ and I think that’s an important conversation right now because there’s a lot of AI-generated content, but being accountable for it is something that I think is just required for us to make it into the next phase.”