IBM Uses AI for the Next Generation of Support

This month, IBM had an interesting briefing on how AI could be used to improve the customer support experience, drive deeper engagement with customers, and effectively improve customer loyalty while dramatically reducing support costs. This sounds unusually good given how new AI seems to be, but IBM has been working on Watson, now watsonx, for decades with AI experience and knowledge that rivals the king of AI, NVIDIA (with whom IBM also partners).

IBM’s watsonx remains the only truly mature, secure, and reliable AI in the market for enterprise users, though I expect IBM’s hard-won knowledge about how to properly provision, deploy, manage, and assure an AI implementation gives it a significant market advantage right now.

IBM correctly argued that an AI deployment team is a lot like a pit crew. A lot of things need to be done by experienced people very quickly or the car, or in this case the company, will lose the race. Losing the AI race may not be survivable, making the quality and execution of the deployment team strategically important to the company doing the rollout.

Let’s focus on how AI could significantly improve the support experience for customers if it is done right.

Assuring Accuracy

One of the major issues with the current crop of generative AI platforms is the lack of quality in the answers. AIs have been hallucinating due to improper or inadequate training (by developing training sets off of the public cloud and directed efforts by users to corrupt the AI during its training). Accuracy in numbers goes to the core of who IBM is, and this goes back to when IBM, early in its computer career, rolled out tools for the financial market which requires extremely accurate results.

During the support experience, you don’t want users struggling to find the right answer, and you certainly don’t want them getting wrong answers. So, assuring the quality of the response in a support engagement goes directly to assuring customer satisfaction and loyalty.

Insightful

Much of the problem with tools like Siri is that they just stripped content off the web, did a text-to-voice conversion, and then supplied an answer. Given web content, this was often neither insightful nor accurate. Early on in the development of Watson, IBM focused on the medical community which found it hard to identify diseases and remedies quickly.

The example I was given back then was from one of the medical advisors who had previously worked for over three years to identify a unique illness that was causing a patient a lot of pain and making it so she couldn’t work. After three years of research, he found the disease and was able to suggest a cure that worked. Watson, when given the woman’s symptoms, made a list of illnesses, and the one she had was in the top 10 (it was number 6) which would have allowed her cure to show up years earlier.

You want the AI to answer accurately and completely and provide actionable information. Otherwise, it won’t be able to justify its cost. IBM’s AI solutions are designed to do exactly that.

Personalized

We are all different in personality, interests, where we work, what we do, and the kinds of questions that we will most often need the answer to. AI has to be able to take what it knows about the user to craft the ideal answer because most of us have really poor communications skills. The AI must be able to infer the question when it is mis-asked and still provide an answer that is crafted for the user’s unique needs.

Doing this makes the AI’s answers far more actionable. Information that isn’t actionable is often mostly a waste of time.

Immediate

Time is of the essence, particularly with support efforts during which the user may already be upset because something they need to work on isn’t working. These aren’t the old punch card days when people were willing to wait hours for a response. We want to get away from call center services where service personnel are poorly trained and informed except when it comes to putting people in desperate need on hold.

The faster you can provide an actionable response, the happier the customer is going to be and the more likely they are to leave with a positive impression about the company.

Initiative-taking

This is one of the areas where AI will eventually shine, and IBM is making that “eventually” part more near-term. Ideally, you’d want support to call and help you prevent a problem before it gets serious rather than waiting until you are hard down and unable to work. A professionally trained and connected AI should be able to better anticipate and help you avoid critical problems so they don’t make it up to senior management who might conclude you are the problem that needs solving.

Wrapping Up

When it comes to AI, few, if any, know more about it than IBM, which aggressively started its own path to AI decades ago. IBM knows what users want from a service, particularly when calling in for service and support, and has crafted the watsonx solution that is second to none in terms of its maturing, accuracy, insight, personal nature, speed, and ability to anticipate and help mitigate problems before they become problems.

When it comes to a mature AI platform that can be focused effectively on support, I doubt anyone does it better than IBM.