The technology you use should be guided by the user experience you want to achieve.
how people do the task today. Figure out what’s valuable, and how you can enhance the experience.
Try to explain in plain language what your AI can do, and where its limitations are.
under-promising and over-delivering is a good way to build trust.
AI is only useful if we understand its decisions. Ideally, the user should be able to trace any result back to the supporting data points.
If you aggregate data from multiple sources, break them down to let the user reproduce the result. This information should be available as part of the user flow through a consistent interface.
You could show a percentage, or try a more abstract visualisation (e.g. star ratings, colored indicators). For results that have multiple parts, break down the confidence for each.
When the input is clear and the answer certain, you don’t want the user to hesitate.
Systems Smart Enough To Know When They’re Not Smart Enough
Not everything should be automated. Most tasks have some parts that are a good fit for AI, and ones that should be left to humans.
being able to intervene, provide feedback, reverse bad actions and reward good ones. AI is more empowering when it works with the user, not for the user.
Balance predictability and serendipity Any personalised AI adopts the user’s bias. This is great for tasks that require predictability, where you need consistently effective results. But for other tasks, it limits our curiosity. It constrains us to options inside our comfort zone.
Prototype with real data and fake AI Using real user data for early prototypes helps you build your machine learning model on the right assumptions. You can use the wizard-of-oz method to get the user experience right before actually building the AI.
Data analysts, researchers, developers, marketers and designers all need to work together to build a cohesive product.
Consider open-sourcing the AI of systems that make critical decisions. Sharing insights with your users and the community builds trust and goodwill.
The technology you use should be guided by the user experience you want to achieve.
how people do the task today. Figure out what’s valuable, and how you can enhance the experience.
Try to explain in plain language what your AI can do, and where its limitations are.
under-promising and over-delivering is a good way to build trust.
AI is only useful if we understand its decisions. Ideally, the user should be able to trace any result back to the supporting data points.
If you aggregate data from multiple sources, break them down to let the user reproduce the result. This information should be available as part of the user flow through a consistent interface.
You could show a percentage, or try a more abstract visualisation (e.g. star ratings, colored indicators). For results that have multiple parts, break down the confidence for each.
When the input is clear and the answer certain, you don’t want the user to hesitate.
Systems Smart Enough To Know When They’re Not Smart Enough
Not everything should be automated. Most tasks have some parts that are a good fit for AI, and ones that should be left to humans.
being able to intervene, provide feedback, reverse bad actions and reward good ones. AI is more empowering when it works with the user, not for the user.
Balance predictability and serendipity Any personalised AI adopts the user’s bias. This is great for tasks that require predictability, where you need consistently effective results. But for other tasks, it limits our curiosity. It constrains us to options inside our comfort zone.
Prototype with real data and fake AI Using real user data for early prototypes helps you build your machine learning model on the right assumptions. You can use the wizard-of-oz method to get the user experience right before actually building the AI.
Data analysts, researchers, developers, marketers and designers all need to work together to build a cohesive product.
Consider open-sourcing the AI of systems that make critical decisions. Sharing insights with your users and the community builds trust and goodwill.
