There should be zero separation between an enterprise’s business objectives and its AI implementation, according to those with experience implementing artificial intelligence in their organizations.
You have to think about the business problems first before you drive the tools in that direction,” said John Daly, senior vice president of worldwide production services at Sony Pictures Entertainment. “Be crystal clear on where the business needs to go.”
Daly spoke at the AI World Conference & Expo 2017 in Boston. He said that an enterprise’s key performance indicators (KPIs) or higher-level market strategy shouldn’t change just because it starts using powerful new AI tools. Instead, market strategy and KPIs should remain constant. The AI tools should be aimed at improving existing processes, not creating new ones.
Daly co-developed an AI tool from Algomus, which automatically generates data reports and supports natural language queries. Daly’s team uses Algomus to track stock of DVDs and other merchandise at retailers, which helps them know when to send more or scale back deliveries. Daly said because improving retail sales is a core business objective, it makes sense to look for ways AI could improve that process, rather than starting with an AI implementation and then looking for business use cases.
It says a lot about the state of AI tools today. The term AI evokes a futuristic vision where computers can answer any question or perform a limitless array of tasks. But enterprises that have implemented AI tools have seen more limited value. It’s not that the tools aren’t useful for some — it’s more that the hype has outpaced the reality. Improvements from an AI implementation are incremental, rather than transformative.
A big part of realizing the desired business impact, Gupta said, is making sure lines of business are involved in implementation and projection planning. “It’s critical to bring those business users to the forefront so that they’re solving business problems,” she said.
Gupta said Monsanto is currently working on about 50 deep learning projects, which include discovering new ways to make crops resistant to diseases, for example. The number of projects grew organically, rather than as part of a forced initiative. Business users at Monsanto who were engaged in early-stage projects that eventually panned out promoted their successes and evangelized the technology, increasing demand for it.
Getting the business involved “helps ignite throughout the company an interest in doing these things,” Gupta said.
Read the source article at TechTarget.