For the digital craftsman and the studio head, the current AI boom is less a revolution and more a balancing act between creative velocity and the bottom line. While the promise of generative tools to draft code, analyse feedback, and compose copy is undeniable, the financial reality is shifting rapidly. At 8×8, a communications platform provider, the narrative is currently one of net savings rather than runaway expense. Their finance team reports that the cost of running Anthropic’s Claude is significantly outweighed by the money saved cancelling dozens of redundant software subscriptions. Joel Neeb, the company’s chief transformation and business operations officer, notes that while the bill is “well below” the estimated $5 million in annual savings, that gap is expected to narrow as adoption deepens and tasks become more complex.
The Token Economy
However, 8×8 is an outlier in a market increasingly obsessed with “tokenomics”-the management of the soaring costs associated with AI usage. Tokens represent the fundamental unit of content an AI model processes, and as companies pour hundreds of millions into these tools, the bill is becoming a central strategic concern.
The pressure is visible across the sector. Last month, the CEO of Royal Bank of Canada revealed a 500 percent surge in token usage over six months. At Cisco, daily adoption is so high that CEO Chuck Robbins described the consumption as “pretty, pretty crazy.” Spenser Skates, CEO of analytics firm Amplitude, reported engineers spending thousands of dollars monthly on tokens. Meanwhile, Box CEO Aaron Levine has noted that budgeting for tokens has become one of the most heated topics in executive discussions.
Data from AlphaStreet indicates a sharp rise in scrutiny. In April and May of this year, roughly 300 companies addressed token concerns during earnings calls or analyst discussions. Compare this to just 93 mentions a year prior, and the trend is clear: the industry is moving from experimentation to fiscal accountability. Executives are now scrambling to implement monitoring systems that select the lowest-priced models for specific prompts or deciding whether to hire more staff or simply increase the token budget.
Spending Without Guilt
Despite the broader anxiety, some firms are doubling down on usage before the tab becomes a problem. In April, Baseball Lifestyle 101, a clothing brand expecting $250 million in sales, instructed about 50 top managers to spend the equivalent of 20 percent of their monthly salary on AI tokens. Bill Rom, the company’s cofounder and chief strategy officer, estimates costs could exceed $100,000 a month by year-end, yet the return is immediate. Claude recently helped secure a $1 million order by identifying a stock shortage in ice-cream-patterned shorts-a task that might have taken days now takes an hour. Rom argues it is vital to inspire usage before enforcing financial constraints.
At 8×8, the approach is slightly different but equally focused on adoption. All 1,800 full-time employees are encouraged to monitor a dashboard showing their and their colleagues’ usage. The goal is not punishment, but collective progress. “It’s not punitive in the least; it’s really just so that we all stay tightly packed in this journey,” Neeb says. Currently, product and customer success teams are the heaviest users, while sales and finance lag behind.
Managing the Laggards
Neeb acknowledges that caps may eventually be necessary. He recently discussed the idea with the CFO due to the rising cost of the new Claude Opus 4.8 model, which costs nearly 1.7 times more than the version released in February. Future access might require proving that older, cheaper models cannot complete the task. “Can we downgrade the model a little bit and still get the same outcome?” he asks.
Despite potential restrictions, 8×8 is not retreating. Customer satisfaction metrics are trending higher, and revenue has grown for four consecutive quarters, with Neeb suspecting a link to AI-accelerated sales analysis. The company’s stance remains firm: employees must gain AI fluency. “If you’re not using AI in some capacity for your role, then you’re missing the opportunity to go faster and get better answers more effectively than your peers,” Neeb states.
This mirrors mixed results elsewhere. At Amazon and Meta, some workers use AI merely because they feel forced to, while others slack off because the tools free up their time. Neeb warns against letting AI speed up work only for employees to take longer lunches or sit on the beach. He is specifically targeting “laggards” like the finance team, which accounts for 28 percent of the workforce but only 15 percent of token consumption. A recent AI hackathon aims to automate manual processes like accounts receivable and quarterly reporting.
Even Neeb himself uses the tool to automate daily emails summarising industry tips. Noticing the high token cost, he asked Claude to rework the automation, cutting usage by 80 percent. The lesson is clear: efficiency must be built into the workflow, not just assumed.
Key takeaways
- While the industry faces rising “tokenomics” costs, 8×8 currently reports that AI savings from cancelled software subscriptions outweigh the operational bill.
- Adoption strategies vary: some companies like Baseball Lifestyle 101 mandate high spending budgets to encourage experimentation, while others like 8×8 monitor usage to prevent waste.
- Executives are actively balancing the need for AI fluency with fiscal responsibility, potentially introducing usage caps or model downgrades as costs escalate.
- The long-term success of AI integration depends on aligning productivity gains with actual revenue growth rather than allowing tools to become sources of inefficiency.
