ChatGPT: the consumer interface to AI goes mass market
What the launch of ChatGPT means for companies and managers - not technologically, but in terms of how expectations around automation will change.
On 30 November 2022 OpenAI launched ChatGPT for the general public. One million users in five days. One hundred million users in two months. No technology product in history had reached that audience that fast.
I do not want to write about how large language models work. That is a long conversation useful to engineers. I want to write about what this launch changes for people who run companies.
The key word here is "interface." Language models existed before ChatGPT. But ChatGPT gave them a user interface that an ordinary employee - not a developer - can use without any instruction.
What actually happened
Before ChatGPT, accessing powerful language models required either technical integration via an API or a subscription to specialised tooling. This meant that using AI at work was a task for the IT team, not the average manager.
ChatGPT removed that barrier. Now any employee can ask in plain conversation for a draft email, a reformulated text, structured thinking, or a document outline - and get a result immediately.
This does not mean AI became "smart" in the conventional sense. A language model works differently from human reasoning and has known limitations. But the usage threshold dropped low enough that it became a matter of individual choice for each employee, not a company-level technical decision.
What changes for managers
The first consequence is a shift in expectations. When a department head sees that a commercial proposal draft can be ready in three minutes instead of three hours, their expectations about work speed change. This applies beyond text tasks.
The second is uncontrolled adoption. Employees will start using ChatGPT regardless of whether the company has a policy about it. This raises questions: what data appears in the prompts? What is the risk of leaking confidential information through public tools?
The third is a productivity split. Employees who master these tools will start completing certain tasks fundamentally faster. This creates both opportunities and new management questions.
What not to expect
Language models are good at text and unreliable with facts. ChatGPT states incorrect information with the same confident tone as correct information. This is called model hallucination, and in late 2022 it is a real limitation that anyone using the tool must understand.
This does not mean the tool is useless. It means it needs to be used where a factual error is not critical, or where the output is always reviewed by a person.
Using ChatGPT to automate complex business processes in 2022 is not yet what enthusiastic articles describe. The tool is good for specific tasks: drafts, summaries, reformulations, structuring. For tasks requiring factual precision or complex logic - caution is needed.
A practical question for a leader
Does your company have a clear view of which employee tasks could be accelerated with tools like this - and what the risks are in doing so?
If the answer is "we have not thought about it yet" - now is a sensible time to start. Not because something needs to be implemented urgently. But because employees have already started using these tools on their own, and it is better to have a considered position than to discover a problem after the fact.