How much AI do I need for my enterprise? What do other enterprises pay for AI? I would like to think that these questions aren’t being asked in today’s marketplace, but I know well that IT professionals are asking these and many others every day.
The issue is that AI isn’t a commodity that one markets like a widget; it’s a tool that can solve business problems. Successful developers, consultants, and salespeople don’t sell AI; they sell solutions to business problems (AI is just one part of it). For perspective, when VoIP came in to focus at the turn of the century, few people had success selling VoIP as a term, “Want to buy some VoIP?” It was sold to reduce infrastructure costs and improve resiliency.
The same principals must be used when marketing AI-based opportunities. Success will come by focusing on specific business problems, and AI-based use cases that will solve them. Here are a couple of examples of how AI can be used to deliver on specific business issues:
Example 1:
Problem – A major hotel chain found that a large percentage of agents were handling non-revenue related inquiries (Do you take pets? I am lost, etc.).
Solution – By implementing a natural language IVR, callers are now answered by a virtual agent. Using AI, caller intent is determined, and a logic program is launched. In the cases for the calls above, the caller would be texted the pet policy or a Google map to the hotel. At the slightest problem, callers are escalated to a live agent. A 30+% cost savings was achieved within the contact center as agents focused primarily on revenue related calls.
Example 2:
Problem – A credit union during busy periods routes simple transactional calls to an outsourced contact center, costing $80,000 per month.
Solution – Like example 1, a natural language IVR determines caller intent (checking balance, recent deposit, etc.) via a virtual agent (VA) and AI. The VA then authenticates the caller and states the account balance or recent deposit. A monthly savings of $50,000 is achieved.
The key to making both sides (enterprise and vendor) of the equation happy is by first stating the business problem and then creating "next steps" to solve it. IT professionals should be leery of any provider that oversimplifies the problem and solution (“we take the call, then through the magic of AI, reduce your costs by X%”). Similarly, a solution shouldn’t be painted as “plug and play.” There will be issues, tweaks, learning, and updates throughout the planning and implementation process. Vendors should break the process down into individual steps, illustrating what technologies and applications are involved at each step.
The better prepared an IT manager is before implementation, the greater the chance of a project being successful and also growing the vendor/enterprise relationship.
Don't miss your opportunity to hear about the latest in AI at Enterprise Connect 2020. Check out AI sessions and the Exhibit Hall, and register this week for our lowest rate. Plus, enter the code NOJITTER at checkout to save an extra $200.
