At H2O World in Mountain View last month, DataScience.com CSO William Merchan joined other experts from Enlitic, Fastdata.io, Kespry, and MapD as they shared how enterprise companies use artificial intelligence and deep learning to transform business operations and strategy. You can watch the full panel, called "AI in Enterprise," below:
During the panel, Merchan shared with the audience how DataScience.com clients have used the DataScience.com Platform to improve efficiency and customer satisfaction.
“We’ve heard that a lot of the roadblocks [to using AI and data science] come with coordinating development efforts with engineering and other stakeholders,” Merchan said. “We help clients remove some of the roadblocks that get in their way.”
One area where DataScience.com clients often use AI, machine learning, and deep learning is in customer service. Platform users have built and deployed models that analyze customer service logs for common themes or issues and then deliver that information to the relevant stakeholders automatically. Ultimately, as workflows like these become more efficient, data science teams are coordinating their efforts more and more.
“We’re seeing a lot of momentum around data science teams now starting to establish best practices internally,” Merchan explained. “At bigger companies, there are even Centers of Excellence being created. This is helping streamline the process and go from first day on the job as a data scientist to being productive.”
For those new to AI and data science, the amount of available information can be overwhelming. “A good starting point is always going to be your internal team,” Merchan said. “Sit down with data engineers internally and understand how the data that you’re consuming internally actually gets created. It may surface opportunities to create new datasets that don’t exist or restructure what you have.
“Start with what you’re trying to influence — whether it is revenue or cost cutting — and work backwards from that,” he continued. “Throwing machines and throwing models and algorithms at a problem isn’t going to solve it. In fact, it may waste a lot of time.”
In order for enterprises to successfully leverage AI, the teams doing that work need to be empowered. Merchan shared insights from his own experiences with the DataScience.com team: “For our team, a lot of it is about individual ownership and applying agile approaches to development. We definitely apply what we preach to our customers in terms of best practices; we want to follow agile development best practices and use a very modern toolset.”