It shouldn’t come as a surprise that many artificial intelligence (AI) projects are failing largely because the processes that organizations are trying to automate are generally not nearly as well understood as most IT and business leaders might think.
To provide some assurance that AI can be applied to a process, Verint Systems this week launched AI Blueprint, a framework through which the IT services firm will first determine whether it is feasible to automate a process using bots and other forms of AI. Once that’s determined, Verint will then build that AI application within a specified time period for a fixed price that starts at $50,000 per engagement.
Tracy Malingo, senior vice president for product strategy at Verint, says Verint is making a concerted effort to reduce the risks associated with investing in AI. Because of all the hype surrounding AI, many business and IT leaders have unrealistic expectations concerning what can be accomplished using AI technologies as they exist today, says Malingo.
“A lot of AI decisions are being made by people sitting on boards,” says Malingo.
The Verint approach calls for spending a month collecting evidence showing how processes really work with an organization and then identifying which of those processes could be augmented using digital assistants, says Malingo.
Verint claims early customers that have implemented the company’s AI Blueprint methodology have already seen a 27 percent reduction in live-chat costs, an 83 percent deflection in live chat volume, and savings of as much as $1 million on customer service email costs in a single year.
AI technologies in all forms can clearly benefit most organizations. The challenge is that simply throwing AI at a process is not likely to succeed. Adopting AI requires organizations to carefully document each step of a process that needs to be automated. The issue that many organizations are now starting to painfully appreciate is that many of the business and IT leaders funding AI projects are far removed from those processes. The truth is, many of those processes are a lot more complex than a bot designed to automate a very narrowly defined set of tasks can handle. Rather than finding that out the hard way, organizations would be well advised to take a second look before making the proverbial leap into the AI void.