A small team applied AI to recurring operational work
OpenAI published a customer case study on September 2, 2026, describing how ATV Big Air Tour uses ChatGPT Work across event listings, merchandise inventory and website analysis. The business operates a national schedule of live ATV and dirt-bike performances under co-founders Derek and Larissa Guetter. OpenAI characterizes its leadership operation as a two-person team coordinating nearly 26 tour dates from May through November. The tour's own website confirms an active 2026 schedule spanning fairs, speedways and other venues, where dates, showtimes, addresses and ticket information must remain synchronized.
The case study is useful because its workflows are concrete. Larissa Guetter handles marketing, branding, websites, media, sponsorships and daily operations, according to the company's founder page. Those responsibilities create repetitive information work around every performance. ChatGPT is not reported as riding vehicles, managing venue safety or making final commercial decisions. It searches for inconsistent listings, organizes information extracted from merchandise photographs, generates planning materials and audits website content. This is a practical form of assistance for a small organization that cannot assign each function to a separate specialist.
Listing checks move from manual search to a scheduled briefing
Event information travels through organizers, ticketing services, local media, chambers of commerce and volunteer networks. A wrong date or venue can inconvenience customers even when the tour's own page is correct. OpenAI reports that Guetter had been manually checking about 30 online publications each day. The process consumed roughly eight hours per week. With a scheduled ChatGPT briefing, the case study says the review now takes about one hour, producing the reported saving of seven hours each week.
The automation checks priority sources, identifies inconsistencies, suggests contacts and drafts correction messages. It can also surface pages the owner did not already know existed. That expands the search beyond a fixed bookmark list and turns scattered observations into a reviewable queue. The human role remains important. A model can mistake an old listing for a current one, confuse similarly named venues or select an unsuitable contact. The publication does not report the number of errors found, the proportion confirmed by Guetter, the accuracy of contact suggestions or how quickly outside publishers made corrections.
The inventory workflow combines vision with structured output
For merchandise planning, Guetter photographed the tour's stock and uploaded the images to ChatGPT Work. OpenAI says the system organized the items, created a spreadsheet and produced a visual inventory website and reorder recommendations in less than 15 minutes. The complete inventory and reorder process reportedly fell from two or three days to two or three hours. That difference is owner-reported and has not been independently timed, but it illustrates how image recognition, classification and document generation can be combined in one workflow.
A photograph can be faster to collect than a manual count sheet when merchandise is spread across boxes, tables or a trailer. The model can propose item names, quantities and categories, then transfer the results into a format the owner can edit. It cannot establish that every garment, size or color was recognized correctly. The case study explicitly says Guetter reviewed the recommendations, made adjustments and sent the final order to the supplier. The deployment therefore assisted inventory preparation and purchasing analysis. It did not autonomously authorize or place the commercial order.
Human review converts model output into business action
The retained review step is not incidental. An incorrect count can produce excess stock, missed sales or unnecessary transportation. Reorder quantities may also depend on upcoming locations, weather, audience size and previous demand, factors that a photograph alone cannot reveal. By treating the generated spreadsheet and recommendations as a draft, the owner can compare them with physical stock and business knowledge before committing money. The useful gain is the reduction of initial sorting and transcription, not the removal of accountable judgment.
This pattern applies to the listing workflow as well. ChatGPT can assemble suspected discrepancies and draft messages, while the owner determines which source is authoritative and whether a correction should be sent. A small business benefits when automation concentrates scattered work into a shorter review period. It becomes vulnerable if speed encourages unexamined output. The case study describes a human-supervised process, but it does not publish error logs, correction rates or purchasing differences that would show how much review was required before the results became usable.
Search hits measure machine access, not customers
ATV Big Air Tour also uses a daily website audit aimed at answer-engine visibility. The system examines whether dates, locations, ticket details and frequently asked questions are structured so AI services can retrieve them. One audit reportedly found that ChatGPT could not access about 90 percent of the site's frequently asked questions and suggested a remedy. This type of analysis can help a small company notice technical presentation problems that make accurate information difficult for search tools to extract.
OpenAI reports that OpenAI search and user-bot hits rose from 183 to 2,421 across consecutive 30-day periods, a 1,223 percent increase. Guetter filtered the analytics to exclude training bots and other AI platforms. The denominator and resulting percentage are clear, but the metric is still server traffic associated with OpenAI search or user bots. It is not a count of unique visitors, families attending shows, ticket purchases or merchandise revenue. Bot activity may indicate improved machine discoverability without establishing that more people found correct information or completed a transaction.
Operational gains can improve access to accurate information
The public benefit in this example is modest but real in form. Accurate event listings can reduce wasted travel, missed performances and customer-support requests. A clearer website can help search systems return dates and locations instead of incomplete answers. Faster inventory preparation can give an owner more time for show production, staff coordination or customer communication. These gains matter most to small firms because the same person often manages both strategic and administrative work.
The evidence does not establish how many listing errors were prevented or whether customers experienced fewer problems. It also does not show that the saved hours increased wages, reduced ticket prices or improved business survival. Those broader outcomes would require records beyond the case study. The strongest supported point is that several recurring tasks became shorter according to the owner and were reorganized into reviewable workflows. The technology's value came from handling search, extraction and drafting together while leaving decisions about corrections and orders with the business.
The case study is specific but not comparative
OpenAI is both the product provider and publisher of the account. ATV Big Air Tour supplies the operational experience and measurements, while no independent auditor, comparison business or published task log is identified. Before-and-after recollections can capture large differences, especially a shift from days to hours, but they may not use standardized timing. Workloads may also vary between inventory cycles. The report does not state how many cycles were observed, how merchandise complexity changed or whether the same quality standard was applied before and after adoption.
This limitation does not erase the implementation. It defines what the case can support. A controlled productivity estimate is unnecessary to learn that photographs can seed an inventory draft or that a scheduled search can consolidate listing review. It is necessary, however, before generalizing the exact time savings to other businesses. A retailer with thousands of products, formal procurement controls or several warehouses would face different validation costs. The tour's results are evidence of one operating pattern, not a universal efficiency rate for small companies.
The durable lesson is workflow design
ATV Big Air Tour placed ChatGPT where inputs are abundant and outputs can be checked: public listings can be compared with an authoritative schedule, inventory suggestions can be compared with physical merchandise, and website recommendations can be compared with analytics and retrieval behavior. Each task has a visible object of review. That structure makes errors easier to catch than in work where correctness depends on hidden assumptions or specialized judgment that the reviewer cannot independently inspect.
A two-person business reports recovering meaningful time from repetitive work and gaining tools normally divided among operations, merchandising and digital marketing roles. Human review remained part of the purchase process, and the traffic increase does not prove commercial growth. The advance is nevertheless practical: AI can combine searching, image interpretation and document creation into workflows that small teams can supervise. Its value should be judged by accurate completed work and usable time, not by generated material or bot traffic alone.