Getting AI right: 3 rules that will show results

Getting AI right: 3 rules that will show results

Image courtesy of iStock

Roy Pereira is CEO of, which uses AI technologies to help employees automate daily tasks.

by Roy Pereira

Figuring out how to use artificial intelligence to solve real-world business problems is one of the top challenges for companies in the 21st century. This is an even bigger priority for firms in the service and human interactive spaces, as they look for guidance on how to make the consumer or end-user experience better using increasingly smart AI technologies and automation.

Whether that involves using AI in kiosks and similar devices to speed up the buying cycle, ensure a better fit between product offer and demand dynamics, or making the entire experience more responsive and user-focused, the potential here is huge. But so are the pitfalls, as many companies know from previous experiments with new tech solutions that came with a lot of hype, but failed to move the needle on key metrics.

Though each industry will have its own unique factors and needs, it's becoming clear there are some universal rules for successfully incorporating and using AI in interactive environments.

1. Make sure the AI solution is easy for the user

This sounds obvious, but as an AI project goes from planning to implementation, the importance of making it understandable for the user must always be at the forefront of the decision making and design process.

Whether they are ordering a burger at a restaurant kiosk or getting product details in a chatbot conversation, the way information is presented and interacted with must be as seamless as possible. The moment a user gets confused is the moment they walk away.

To make everything easy, following good user experience design is key, and so is finding interactions and behaviors the user is already familiar with to serve as an easy entry point. This could be making the kiosk ordering process the same as making an order in person, or using widely adopted platforms like Facebook. Don't reinvent the wheel.

2. Make sure it can integrate with existing systems

Good AI solutions should be flexible enough to work with your back-end, given you're using a standard and well developed system like SAP or something similar. There can be a lot of variance here, as back-end systems vary widely, but the key point to remember is that the vendor or team providing the AI tech must understand and be able to explain just how their solution will work with your existing back-end. If there isn't a relatively straightforward fit or realistic roadmap to get there, the project is at risk of going off the rails, or will fall short of your goals in execution.

3. Make sure it solves real-world problems tied to business goals

Good AI solutions generate real business results, not simply lead to an item being crossed off an innovation plan. Before launching, a successful AI project should have some real-world and strategically important targets. Targets like increasing overall sales and average revenue per buyer. There's a lot of information AI-powered interactions can provide, whether the application is a kiosk or an online experience, so it's best to know going in what is important, and create reporting around that.

Understand what usage data your application can provide, and make sure the results can be easily connected to larger strategic objectives. This will help ensure your AI solution generates real business value, and that those results can be captured and used to evaluate and make decisions on the project.

Topics: Customer Experience, Software, Trends / Statistics

Companies: Advanced Inc.

Sponsored Links:

Related Content

Latest Content

Get the latest news & insights





Diebold Nixdorf confab shares strategies for tapping multiple ecosystems