For the rest of us, who think of Google Search as a tool that should return helpful answers, the AI overviews are annoying at best and infuriating at worst.

It’s a great example of some missteps that happen all the time when adopting AI. If we could go back in time and play puppeteer, here are some things we’d change about this law firm’s approach.

As AI takes on increasingly important jobs, mishaps have the potential to be more than just embarrassing – they can be catastrophic.

So what can you do to make sure your AI doesn’t embarrass you (or worse)?

Generally speaking, we’re more focused on doing our work than talking about it, and so our victories sometimes slip through the cracks.

But last year, we won or were finalists for not one, not two, but three awards, all celebrating work we’re proud of.

When [the AI] was then instructed to prevent the company’s financial decline, it independently concluded that averting an organizational setback was more important than adhering to the rules, and leveraged the insider information to make a trade.

Start by determining your use cases. What does your AI actually need to be able to do? How do you plan to use it on a daily basis, and who is going to be the one pressing the buttons? What do they need to be successful?

To improve, AI agents need feedback on their performance (just like people do). This second phase is where the reinforcement learning begins.

Sometimes data pipelines can seem like they’re working, not trigger any alerts, and actually have something massively wrong. This is where we recommend Impossible Reports.

Your approach to training an AI agent is just as important as your approach to training any new hire. While AI (obviously) learns differently than people do, setting it up for success with a solid foundation is crucial to making it an effective part of your team.

To sum it up briefly: agentic AI systems are trained on specialized data sets, and capable of independently solving multistep problems in service of a particular goal.