Who has the deets about AI replacing the frontline?
8 replies (most recent on top)
Correction: In my long post below, I meant Few Shot Technique (mistyped as "Free" Shot).
Few-Shot Learning (FSL) is a Machine Learning framework that enables a pre-trained model to generalize over new categories of data (that the pre-trained model has not seen during training) using only a few labeled samples (Question and Answers).
enterprise teams would conspire to make AI fail and save jobs in the short term. T-Mobile might need to do another round of layoffs in order to have the right stuff for AI transformation, probably through acquisition. Then install a Chief AI Officer. By the time AI is going to replace Frontline teams, ATT will propose to merge with T-Mobile to compete with Verizon and Amazon. Sci-fi or is this as clear as T-Vision?
MS came to our CEC around spring. Talked about how great our company is doing that others are laying people off, not us (this was right after the second wave of layoffs 6 months prior granted not Frontline) and how soon calls will be reduced because AI will be taking calls.
When that happens don't know. But it sure explains why we have transcripts in memos now. The AI can review those to better assist. See what style of customer support works better for the customer etc.
Once AI is starting to work and gain understanding, the rate it's intelligence works is multiplied rapidly. I too like another commenter was curious. Guy who was helping build AI for Google said they built an arm for this AI. They gave it a shapes puzzle like kids get. Put the triangle in the triangle hole, circle in circle. Once it finally figured out one shape which took a deal of time, the next shape didn't take as long, next shape even faster, next faster then last. This is an example of what I mean by it's ability to gain knowledge. He mentioned AI has been developed to have an IQ greater then Einstein as of now and by next year with be 10 times that level.
So truly the ability is there at this point. And the transcriptions will do nothing but feed the AI with the knowledge of how to handle customer service as well as how to not handle it. Genius really, but terrifying for the workforce/ economy. This just makes the rich richer and poor poorer.
AI isn’t replacing the frontline. virtual retail might.
"Who has the deets about AI replacing the frontline?"
Was this an AI-generated comment? Lol.
Can’t happen fast enough. Have you dealt with frontline lately?
I was let go on the first day 8/24, and as I started exploring the job market, I noticed that almost 80% of the job description I was looking at - had at least one bullet "AI / ML experience is desired/ is a plus etc.
I have very limited insight into what's next at TMO, but what I have heard is that it could be an AI offering from Adobe or from some alliance partner of Adobe. Said that, I started exploring myself as what's the deal with AI behind all the ChatGPT hype.
I can share what I learnt and I will encourage all to explore and get some basic idea of AI (until the hype dies and real substance emerges, we will have to deal with this AI noise by having some basic know-how of what the heck really happens behind the scenes. So this is what I have gathered so far in last few days and my understanding could be off, still sharing what I understood:
These are some of the building blocks of AI -
(1) a pre-trained model (like GPT and others) how they work is mostly proprietary and years of research, not sure how much of the info are in public domain, it is too technical anyways as I gathered
(2) then comes in "Training of Model as per your needs" and that's where Enterprises like TMO are putting the work, so this pre-trained model already knows how to learn, what TMO for example needs to do is - train it with TMO specific stuff, like some prospect coming in, get some initial info and show her/him the best promotions as per the context, real intelligence comes into picture when follow-up questions come from this human, how to respond to those in the context of whole conversation. Like, I could say oh all these promotions are BS, tell me something new.
(3) In continuation of point 2, this is where it is all data, you are continuously training this AI Model with more data to take it to a point where it can start to respond like a human would do.
I spoke to a freelancer as how exactly this is being done, he told me there are "Techniques" to train the AI Model, and more efficient the technique is more complex it is to understand at first. However, he showed me one of the simplest "Technique" out there - Free Shot. Look it up, basically it is feeding Q and A to the model and when you ask a similar question, it respond by picking up an Answer for the similar question you have already trained it on. It gave me lot of confidence to see how exactly it is wroking, after hearing all the overload from Internet as "what happens". Now I know it is not some magic, but lot of work, voluminous work happening behind the scenes. As I see, this voluminous work pays off because once you are done, you can reuse that as many times as you want. Of course when things change, your AI Model will need new training but my guess is for changes, it is lot less work to do.
So I would say, start exploring AI and easiest way to get started is talk to people who are really working in this field and know what they are talking about and are not just "Noise on the Web".
Sorry for the looooong comment, slightly off topic too !!
I am not excited about that. These virtual chat bots, SMS agents, etc are a pain in the butt. You get in a weird loop and can't actually talk to anyone or get answers. I've had this issue with several other companies so I hope they test the heck out of AI before releasing.