ChatGPT for everyone or...

Is it worth spending 30 USD per employee for a ChatGPT license? My take is no—most employees simply won't tap into its full potential.

1. The Gap Between Potential and Reality

I've been launching pilot projects that include ChatGPT, and the reality I've observed is that many people either don't use it at all or only use it for basic tasks, like a search engine replacement. This reveals a clear gap between the capabilities of ChatGPT and the actual usage patterns within most organizations.

2. The Ideal User Profile

I started analyzing the ideal user profile: someone who knows how to collaborate effectively via chat/email (therefore knows how to write in a clear and concise way), has a strong technical background in IT, is also a domain expert in some areas, can explain work in both broad strokes and fine details, and is shameless about experimenting—but still has a clear purpose beyond the technical "challenge."

Turns out, not many people have this combination of skills—and that's why investing in large-scale licenses for a "holistic co-pilot" that most people won't effectively use is a real concern.

3. Focus on Specialized Tools

A better solution? Focus on specialized tools tailored to specific, deep-dive problems—like accounting, coding, translation, etc.—that are integrated seamlessly. Tools that handle a problem end-to-end, rather than relying on a general-purpose AI.

4. Deploy ChatGPT to Champions

Ideally, ChatGPT should still be deployed—but to a small team of champions, equipped with API access. Let these champions prototype and execute new ideas, and then turn successful experiments into use cases for the broader organization.

5. Alternatives to Full Licensing

There are also alternatives that mimic ChatGPT using APIs. These approaches can boost productivity on a large scale without breaking the bank. By setting up customized AI solutions that are tailored specifically to an organization's needs, companies can gain the benefits of conversational AI without committing to expensive, per-employee licenses.

6. The Path Forward: Incremental Gains

We may eventually reach a point where everyone is "plugged in" to a ChatGPT-like system, seamlessly collaborating with AI to draft reports, automate workflows, or even manage projects, and working incredibly efficiently—but we're not there yet. This kind of future will require not just better technology, but also a shift in how people work and think about AI.

Until then, it's better to take baby steps, iterate, and aim for incremental gains. If your organization improves productivity by just 1% each month, the cumulative value will be immense over time.

I am curious if my observations on three big companies, distributed worldwide and in very different domains, are correct.