What People Actually Use ChatGPT For (And Why It Matters for Small Business)

The most useful part of the OpenAI usage study is not a flashy headline. It is the shape of real behavior.

The paper tracks ChatGPT usage from launch through July 2025 and reports scale that would have sounded unrealistic a couple of years ago: around 700 million weekly users, roughly 18 billion messages per week, and adoption at around 10% of the global adult population.

Those numbers matter, but the category breakdown matters more.

The largest conversation topics are practical guidance, seeking information, and writing, together making up close to 80% of usage. Technical help exists, but it is a much smaller slice than most "AI discourse" implies.

That is exactly what your chart visualises. The mainstream use case is not "replace developers". It is "help me think, decide, and communicate better right now."

For small businesses, that should change investment priorities.

This is a decision-support story first

A lot of AI conversations in the market still default to extreme narratives. Either full automation fantasy, or complete dismissal.

The usage data points to a more practical middle.

People are using ChatGPT as a decision-support layer inside normal work and life flows. They are asking for guidance, interpretation, drafting help, and structured thinking support. They are not waiting for a perfect autonomous system before using it.

In business terms, that means AI value appears first where communication bottlenecks already exist.

It shows up in faster briefs, clearer updates, cleaner handovers, better first drafts, and shorter loops between "we should do this" and "this is now done."

That may sound less dramatic than autonomous agents. It is usually where the money is.

Why smaller businesses should care more than they currently do

Small teams are disproportionately affected by communication drag.

When one person is blocked waiting for clarity from another person, a whole workflow can stall. In larger organisations, that delay can be absorbed. In smaller ones, it is usually felt immediately in delivery pace, responsiveness, and client confidence.

This is why the writing and guidance-heavy usage pattern is commercially relevant.

If AI helps your team produce clearer output faster, you are not "just improving writing." You are reducing cycle time on core operational decisions.

That creates practical upside in places owners care about:

• faster follow-up on opportunities,
• fewer dropped handoffs,
• clearer delivery communication,
• and less founder time spent translating rough internal thinking into client-safe output.

The paper also shows an important maturity signal

The study notes non-work usage has grown from about 53% to more than 70% of usage over time.

That does not mean workplace value is weak. It means these tools are becoming general-purpose cognitive utilities across contexts.

For operators, this is a maturity signal.

When a tool becomes this embedded, the question is no longer "should we try AI?" The better question is "where does this remove friction in our existing workflows without creating new risk?"

That framing produces better outcomes than tool-hunting.

What to do with this as a small business owner

Do not begin with a big transformation programme.

Begin where your team repeatedly loses momentum because someone needs to explain, summarise, draft, or structure information and it keeps slipping to "later."

Then apply a disciplined sequence:

Choose one workflow with clear commercial importance. Improve quality and speed with AI support. Measure whether cycle time and output quality improve. Keep what works. Drop what does not. Repeat.

This sounds obvious, but most teams skip the measurement step and end up with scattered AI usage and weak accountability.

The businesses that compound value are the ones that turn AI from ad-hoc prompting into operational standards.

Bottom line

The OpenAI/NBER evidence supports a pragmatic view.

ChatGPT is being used at scale primarily for guidance, information, and writing. That is not a side note. It is the core adoption signal.

For smaller businesses, the implication is straightforward: the first wave of value is usually not deep technical automation. It is better decisions, faster communication, and tighter execution in the workflows you already run every day.

That is where practical ROI appears fastest.

---

Primary source paper: How People Use ChatGPT (NBER Working Paper 34255, September 2025) https://www.nber.org/papers/w34255

Direct paper PDF: https://www.nber.org/system/files/working_papers/w34255/w34255.pdf

Hero image source: user-provided chart screenshot in Birdcage workflow context.