Tag Archives: ATR

So, I was attending a setting where I was dependent on customer support. I am being intentionally vague here, because the youthful young sprout side nothing wrong. She was just young (about my age millennia ago), but before you start judging her, and she did nothing wrong. I got to think “What if we had customer support settings on a DML level?” Consider this (overly exaggerated example) 

Tech help: Good morning, how can I assist you?
Customer: My house is on fire
Tech help: Please take a moment to assess how we can help, how can we assist?
Customer: It’s freakin hot, my house is on fire

As you might expect, this will not go anywhere useful soon. And you might think that this is an exaggeration, but when you are assisting a customer and he cannot get to his data, his life figuratively ends. So we need to get tech hardware to assist us. So what if we had a voice measuring setting? Not to interpret (even thought this might help too), a setting where the voice can be measured for stress levels. The technology exist, but consider that this might be overly expensive. How can this technology be made cheaper? Now consider a DML engine that parses the stress level and considers alternative responses so instead of “How can we asses you problem”, state “let us assist you too get your data to you”, but the system gives the tech help options, and the setting gives the 4-5 responses in colour. Red would be ‘Don’t do that’, but others might become options, Orange, Yellow and Green would be available. There will be moments when the Orange is the only one that makes sense, but Green would optionally be the best. And the learning setting that a DML/LLM support system has is that it can keep track of the answers and how it affected the customer. You see, I have been in tech support for decades and there are a few handles you can apply, but the scripted answer is never a great option (I never showed that to my bosses), they had too tender an ego to risk it. 

So when this system would be deployed, optionally with bells and whistles like zendesk, but most of these products are about recording data, not a setting that actively supports the Helpdesk to record and adjust scripts for aiding the tech support. And even if all these AI systems are fake AI, the data for customer service, customer care and technical support exists, there is plenty of it. So these systems are fake AI, but it is based on DML/LLM systems and they could bring a much larger change in this field. And as I see it, change will be required soon enough. The old guard of these systems are retiring and the new generation mostly lack experience. So why not let the DML/LLM system tweak the system? 

These were just a few settings I was looking at and at present I have no idea what there is, so I ned to look into this, because I might have developed an idea here, but perhaps so did someone else and I need to look into this to see what there is. So it is a little out of the blue, but I have been involved with customer care and technical support for decades, so I might has an idea or two to help this along. So, I need to mull over a few things and as I had nothing to offer DARPA (they are all in drone mode) I need to find a new hobby in the non-drone setting. Although destroying the Iranian railway systems are done based (as was my handle to destroy their refineries). And as DARPA is in delusional mode (as I personally see this) to get a drone carry twice the weight of the drone, is simply ridiculous. The Cessna 408 SkyCourier couldn’t do it, the ATR 72-600F couldn’t do it and the BAE 146-200QT couldn’t do this, and they have people in place who tried that for a life time, so why push the cogs of a civilian setting? I felt pretty proud that I as a non-expert in drones found a way to destroy Iranian railway lines and refineries. But I do believe that DARPA is taking this to a delusional stage. 

Still for now lets see what we can do to improve customer carer lives and reduce the stress they are confronted with. A much more rewarding result. Don’t you think so?

Have a great day.