It started with AI slop. It then became Work slop¹. Are we in danger of another onslaught. I’m faced with an abundance of fully fledged Apps, Agents and tutorials about building them, by anyone. This is on top of an already bloated App marketplace. Furthermore, the Apps or Agents are dense with functionality that is mostly unusable or just plain confusing and wasteful. The trend is more is better, when it should be less.

On the face of it

When you see the output of something that has been built by AI using the power of language (prompts), it’s incredible what is possible. You are blown away. If you’re like me, a techno geek, its hard to not be impressed. Even to get carried away.

You marvel at the possibilities. You get caught up in the eye candy that is every button and knob, or graph if they are included as they so often are – especially where analysing data is concerned.

Only when you start using does it become evident that there is more than meets the eye. And therein lies the rub.

No usability testing has typically been performed by the time it is blithely shared with the expectation that you MUST adopt it, and it’s the next big thing.

There’s no regard for the user and all the other tools he or she needs to contend with. No consideration of the fact that there may be a dozen competing tools out there already the user needs to consider. Or if it is new, unique and useful, that the user will need some kind of onboarding, instruction or adoption support.

Don’t make me think

Don’t Make Me Think: A Common Sense Approach to Web Usability by Steve Krug was originally published in 2000. I think we need to come back to this sensible way of thinking.

When we are outsourcing so much of our thinking to AI, let’s not throw the baby out with the bathwater.

If you are going to build something, don’t just share the first iteration. Use AI to analyse usability and refine the output before just throwing out a solution as people often do.

The risk is everyone becomes overwhelmed and true value is lost if you don’t stop, think and be critical first.

Discernment and brutal prioritisation

You don’t have to be a designer, developer and/or software usability tester to be able to arrive at the right decisions. Just use common sense and critical thinking.

Discernment about what an App or Agent should do and with the number of features it should do it with, is where you start. Think like a minimalist, remove until it breaks. What are the minimal amount of capabilities needed to get what you think should be done, done. Refine to the lowest common denominator and get to those features and functions before shipping. All it takes is an instruction.

What you think an App or Agent will help you and others achieve is where you prioritise. I’m not suggesting where you start but this step is a good option. Before you whittle down an App or Agent to it’s essential and useful features and functions, you may already conclude there is no discernible point to it. Or you may arrive at this stage after making the refinements.

Regardless of where you start and arrive at, think first before just throwing something out there. We’ll all be better off for it.

1. AI slop and work slop both describe low-effort, AI-generated content where the creator puts in near-zero effort while forcing the consumer or recipient to do heavy cognitive labour. However, they differ in their scope, intent, and where they appear. ↩︎