What is ChatGPT?

ChatGPT is OpenAI's conversational assistant launched in November 2022: the product that turned generative artificial intelligence into a dinner-table term, not just an academic paper's subject. It popularized chat-based interaction with a language model for writing, summarizing, coding, translating and reasoning through a problem, and it is today among the software products with the fastest user base growth ever recorded. Under the hood runs OpenAI's GPT model family, but the product has grown well beyond text chat: conversation memory, file and image upload, code execution, web search and third-party tool integrations. For a company the distinction that really matters is between the personal consumer account and the Team or Enterprise plans, with different data agreements and no training on business content: it is precisely that distinction, often overlooked by the people using the tool, that sits at the heart of the shadow AI problem.

What the tool covers

Under the hood runs OpenAI's GPT model family, but the ChatGPT product is more than the models: conversation memory, file and image upload, code execution, web search, third-party tool integrations. For companies, what matters most is the distinction between the consumer plan (personal account, data that by default can contribute to training) and the Team/Enterprise plans (different data agreements, no training on business content, admin controls). It is a distinction the people using the tool often do not know, and it sits at the heart of the shadow AI problem: employees pasting company data into a personal ChatGPT account, in good faith, without knowing which plan they are on.

How to evaluate it as a buyer

ChatGPT remains the reference point for widespread adoption and a near-zero learning curve: it is often the entry door to GenAI for an organization. The real work for a company is not deciding whether to ban it, but governing it: knowing who uses it, with which account, on which data, and offering an enterprise alternative (Team/Enterprise or another vendor) before shadow AI becomes the de facto norm. It should be evaluated as a peer of the other general-purpose assistants (Gemini, Copilot, Claude): the choice depends on the ecosystem the company already lives in, not on any presumed permanent technical edge. Picking the right tool is only the first step: if the project then stalls before reaching production, the problem is almost always somewhere else, as I cover in the post on why AI projects stay stuck in pilot mode.