OpenAI’s ChatGPT Enterprise is designed to bring its AI assistant into business settings, where teams may need to work with longer documents, manage access across an organization and address data security concerns. The offering also raises a practical question: can an AI assistant become a routine work tool, and where does it still need human oversight?
More room for longer work
Enterprise customers get unlimited GPT-4 queries at faster speeds, along with a 32,000 token context window. That larger window lets the assistant process more text or sustain longer conversations at once. The source estimates that 32,000 tokens are roughly equivalent to about 24,000 words.
For comparison, ChatGPT with GPT-4 maxes out at 8,000 tokens for ChatGPT Plus users. A longer context can be useful when a task depends on material the user supplies, such as a lengthy document or several pieces of information that need to be considered together.
These capabilities describe what the service can handle, but they do not guarantee that every answer will be correct. The article points to analysis, explanation, summary and translation as areas where the model’s strengths appear to lie, while warning that large language models can make things up or draw false conclusions.
Company data and account controls
Because ChatGPT processes data in the cloud on OpenAI’s servers, security and privacy are central to the enterprise offer. OpenAI says the service encrypts data at rest using AES 256 and data in transit using TLS 1.2+. It also says that customer prompts and company data are not used to train OpenAI models.
The article contrasts that policy with the free and Plus versions of ChatGPT, where OpenAI uses data for training unless conversation history is turned off. For businesses weighing the service, that difference is relevant to how employees handle company information.
ChatGPT Enterprise is SOC 2 compliant. The framework, set by the American Institute of Certified Public Accountants, covers audits of a company’s controls for managing customer data, including security, availability, processing integrity and privacy.
Administrators also get a console for bulk member management, domain verification and single sign-on (SSO), plus usage insights for larger deployments. Those functions give organizations ways to manage access and monitor adoption across a team.
Adoption does not settle the value question
OpenAI says ChatGPT is already used in over 80 percent of Fortune 500 companies, based on an analysis of registered email address domains. It named Block, Canva, Carlyle, The Estée Lauder Companies, PwC and Zapier as early adopters of ChatGPT Enterprise.
OpenAI said those companies used the service for clearer communications, coding tasks, exploring complex business questions and creative work. In a promotional blog, Jorge Zuniga of Asana said ChatGPT Enterprise had cut research time by an average of an hour per day for his team, while helping them test hypotheses and improve internal systems.
Those examples suggest a range of possible uses, but they do not show that the tool will be equally helpful in every organization or task. The article frames its usefulness as dependent on the application. Andrej Karpathy, an OpenAI employee and influential AI scientist, argued on X that organizations should consider AI assistants a basic work tool, comparing them to a new kind of spreadsheet.
Useful assistance still needs facts
The risk of inaccurate output shapes how businesses may want to use the system. Rather than relying on GPT-4 as a factual reference, teams can provide their own facts or data for it to analyze. This approach gives the assistant material to work with while leaving users responsible for checking the result.
ChatGPT Enterprise also enters a market where Microsoft offers Bing Chat Enterprise, based on GPT-4 and other technology licensed from OpenAI. The article said Bing Chat Enterprise was included with Microsoft 365 for Business Standard ($12.50 per user per month) and Premium Plans ($22 per user per month). Microsoft had also said the service would eventually be available standalone as a $5/month feature.
OpenAI had not publicly announced ChatGPT Enterprise’s price. A representative said the cost would depend on each company’s use case and directed interested businesses to contact OpenAI. Without a public price, organizations cannot judge the offer on features alone; its appeal will also depend on their needs and the cost they are quoted.