By the twenty-third week of 2026, the frenzy around tokenmaxxing had started to fade from corporate offices and remote work environments alike. Earlier in the year, businesses had encouraged employees to push AI models to the max—crafting lengthy texts, deep data analysis, and generating complex reports with no limits. But the enthusiasm is cooling down as companies reassess the true cost of using AI at full capacity. June Wang of Bain & Company summed it up clearly: Not every task demands the most advanced and costly AI models. As it turns out, running AI at its highest setting proves too expensive to be practical for many businesses.
Each AI token—roughly representing three-quarters of a word—adds to the bill. What businesses initially saw as a low-cost monthly plan for ChatGPT or other AI platforms soon revealed a hidden cost. These expenses spike when the models process internal data, often through API requests that accumulate with every user prompt. Some organizations found their AI costs ballooning into the hundreds or even thousands per month. This shift has prompted companies to adopt smarter, more cost-effective strategies, like model routing, which sends queries to cheaper, yet efficient, AI models.
Nonprofits leverage AI with fresh tools and training programs
Simultaneously, nonprofits are discovering new ways to harness AI through Google’s updated Workspace for Nonprofits. The company recently unveiled a package of tools and resources, including access to Gemini, a robust AI model, and NotebookLM, a research tool. Nonprofits also gain access to Google applications such as Docs, Sheets, and Meet, all at a discount. Google also offers free AI courses tailored to nonprofit needs, such as fundraising and operations.
For instance, one eligible nonprofit can now use free training to integrate Gemini into their daily operations. The tools allow teams to automate donor inquiries, monitor social media mentions, and produce reports that previously required hours of manual labor. For groups operating with limited budgets, the efficiency and time saved are significant. In their research, Google reported that 86% of surveyed nonprofits believe AI can make their work more effective. The personalized training ensures that even small teams can quickly adopt these technologies, enhancing their impact without adding unnecessary overhead.
Yelp’s AI phone assistant processes over a million restaurant calls
Yelp has recently enhanced its AI phone assistant, Host, making it a valuable asset for small restaurants struggling with labor shortages. The tool can handle incoming calls around the clock, take pick-up orders conversationally, and transfer them directly to a restaurant’s point-of-sale system, eliminating the need for staff to manually process calls during busy periods. The system is designed to mimic human conversation while handling multiple tasks simultaneously, making it ideal for high-traffic environments like cafes and bistros.
One restaurant noted a 38% increase in call volume after using the service. Host also supports integration with OpenTable, allowing it to manage reservations, cancellations, and modifications anytime. According to Yelp, the assistant now processes over a million calls each month. Yelp positions it as a tool for augmentation rather than replacement—ensuring that AI handles routine tasks while allowing humans to tackle more complicated requests. This feature is especially useful for calls requiring nuanced responses or emotional intelligence, which AI systems are not yet fully equipped to handle.
For small restaurants, the cost savings can be substantial. Hiring a full-time server or host to handle phone calls can be costly, often amounting to thousands per month. AI systems like Yelp Host and Slang.ai offer a scalable and efficient alternative. These tools allow users to transfer to a human if the request is complex, preserving the value of in-person service where it’s most needed. Yelp claims that their system has helped several small businesses avoid staffing shortages and reduce operational costs, making it a popular choice among independent operators.
The take-home message for small businesses is clear: AI functions best when it is treated as a practical tool, not a magical solution. Whether it’s streamlining restaurant communications or supporting nonprofit operations, the real impact comes from balancing the cost with the value provided. As companies and organizations adapt to the evolving landscape of AI, the focus is shifting toward efficiency, affordability, and targeted implementation.
While high token usage continues in some areas, the widespread tokenmaxxing trend is declining. The new direction in AI usage is smarter, more strategic deployment—selecting the appropriate model for the task at hand, rather than defaulting to the most expensive one. For many organizations, this means carefully analyzing which AI tools truly deliver a return on investment and which are better left for simpler, less costly alternatives. As businesses continue to refine their approaches, the goal is to maximize output while minimizing unnecessary expenses.