Will AI Be Big at InfoComm?

artificial intelligenceWe are on the eve of InfoComm, and one of my big questions is this: Will AI be a star of the show, ignored during the show or will it be an imposter? I have been, and continue to be, bullish on the power that AI can have on our industry. It is only limited by our imagination. Yet, I will be watching carefully to see if companies are truly pushing AI at the show or if they are rolling out new programming and selling it as AI.

I recently attended a conference put on by a software company we use at my institution. AI, and in particular conversational AI (CAI), was a major driver of the sessions. I really like this company, so I am not going to use their name, as I don’t want to make it sound as if I think they are doing something wrong. However, many of the presenters used the terms CAI and AI completely incorrectly.

Let me start by giving an example of the correct usage of the terms. The company is looking forward to implementing AI in its ticketing product. It will use AI to read through all the current tickets and determine if there is a common issue within them. This would give an IT department a heads-up that something is beginning to go wrong. This rollout requires processing knowledge, applying that knowledge, and then generating new knowledge. This is AI, and it is powerful.

The company is also taking it a step further to then read through responses by both the technicians and the customers to see if there are common fixes. If so, the software will generate a knowledge-based article on how to solve the problem. This is good stuff. The next step in this process is that the software will note when a technician is responding to a ticket, and it will point this person to the article that the AI created. Wow, this is powerful stuff.

Finally, after enough iterations, the system learned the common language around this problem and edited the KB based on feedback in the tickets. It eventually starts responding to these tickets before a technician even sees them. This is AI: It serves our customers better, makes our processes more efficient and allows our technicians to work on even more complicated problems. This is a value added.

Let’s return back to the problem I mentioned at the beginning of this article. The problems I saw at the conference came largely with reference to CAI. Several presenters showed their CAI systems and then showed us what was going on behind the curtains. The systems were not anything near AI. One example was that a person could put in a ticket for a broken iPad screen. Because everyone will write this in a different way (utterances in AI lingo), the system needs to be able to understand these different utterances. People were developing systems in which they programmed 17 different utterances; it was crazy. The computer was not using AI to understand this; someone was programming this into the system. The next step was the “AI” response. Again, this is not an AI response but a pre-programmed response. Each step went this same way. A person spent time attempting to determine what the utterance was supposed to mean (intent in AI lingo) and then wrote a response. Not only was this not an example of AI, it is not even new; similar things have existed on the web for years and years.

See related  Disguise Launches an Entry-Level Media Server Lineup Dubbed EX

The point I am working toward: We need to be critical thinkers and consumers when people use buzzwords like AI. Not only were some of these systems I described above not AI, but one could reasonably argue that between the cost of the programming software to write these systems (it is expensive) and the time spent to program them, they are not providing any type of ROI. If you see a company at InfoComm talking about AI, or claiming it has AI in its systems, be that critical thinker. Ask questions about what it is actually going to do and how it is actually going to improve services to your business or to your customers. The other thing to keep in mind is that true AI is still very (very!) expensive. It is expensive enough that most organizations will not be able to afford it or will need to use a shared system with others. If someone sells you something that seems to be too inexpensive and too good to be true, it likely is. With all that said, I am really hoping for some very creative examples of how AI can move our industry forward and give us even better experiences.