800.518.9750       info@maventanalytics.com

Picture of Newsroom


Five Tips to Get Started With Artificial Intelligence in Your Business

In a recent webinar with Pyramid Analytics, Mavent Analytics Founder and CEO Tony Lopykinski offered five tips for getting started with artificial intelligence.

1. Identify business opportunities

“It really comes down to understanding where your business opportunities are,” Tony said. “Just because you have a hammer, you don’t go around the organization looking for nails. It doesn’t work that way.” Instead, he said it’s about bringing together people with domain knowledge to partner with business stakeholders to solve real problems.

2. Start small

He suggested starting with a “proof of value,” which differs from a proof of concept in that it’s not just about solving a problem, it’s about measuring the value of the solution. “It’s one thing to create a solution and make sure that solution solves the problem,” he said. “It’s another thing to make sure: does it deliver the value? Define those measures for success when you start. Does the proof of value actually deliver on those value measures?”

3. Fail fast; move on

“Adoption is everything,” Tony said. If people aren’t willing to take the outputs of that model and drive real decisions from it and adopt it in their business practices, don’t go forward. Fail fast and move on to the next one and find that one problem that the organization really wants to solve.”

4. Don’t put technology first

“Don’t make it about technologies,” he cautioned. “I always advocate less is more, especially when you’re getting started. Sometimes the most simple models deliver the best results. Sometimes the problem only requires a regression analysis.; it doesn’t require a neural network to solve. Find the problem first and then focus on the solution.”

5. Get the data right

“You can’t implement analytics and AI without having your data house in order first,” he suggested. “Like anything else, if you don’t have good data, if what you’re feeding to do the model isn’t quality, it’s just not going to work because obviously the outputs are only as good as the inputs.”

To learn more, watch the full webinar, “Transcend the Barriers to AI-driven Analytics” here

Talk to a Mavent expert today

Show Buttons
Hide Buttons