If your company is investing in AI you MUST pay attention to this. The abandonment rate on gen AI tools is huge. Why? And how do you fix it? A lot of people's first experience with tools like ChatGPT looks like this: - Think about what to do with it - Find a use case that may or may not be a good fit for AI - Write an awful, generic prompt - Get bad results - Get frustrated and leave ⚠️ This means your company is wasting money on tools that COULD give your team's superpowers (research shows >40% increase in quality for knowledge workers), but people don't get them because they don't know how to use them! We've seen this over and over with our clients: even just a short, well-designed workshop can get your team 10x more comfortable and productive with. You can buy all the tech you want. Adoption won't take care of itself.
100%... the lack of CONTEXT provided to an LLM is ridiculous. As such, it treats the inquirer as the mean and it acts at the mean. That is, an IQ of 100 to be clear. Not so high. You MUST begin with describing who YOU are and who IT should me as deeply as possible. Then prompt. Hard to blame humans though... a few decades of a query-response paradigm has them well-trained.
The thing that perplexes me the most are tools that use AI that are entirely based on a free text prompt. And it happens as it happens with most conversational interfaces: you never know what you can ask, how you have to ask it and how far you can go. While this can somehow work for a generalist tool like ChatGPT, I find it to be a design laziness in many other vertical scenarios where the interface could and should guide the user.
And that’s how Apple managed to leverage chatGPT and integrate it with its own ecosystem with their own conditions
I Totally agreeing with this.
Sami Ben Jamaa, Dr. Eng.
Passionate about the impact of technology on society | Data Governance | Digital Commons
3mograph source?