XtendOps

Behind every great team is a great leader! What a way to end the week! XtraOrdinary Results from XtraOrdinary People with XtraOrdinary Fun! Random snaps from XO CS WEEK 2024. #StayWild Happy CX Week 2024 XO Iloilo!

Remember: Automation is the journey not the destination! ??

POV: You’re rolling out your CX AI strategy, but you’re not sure if you can trust the data… You’re excited—your vendor is sharing some very encouraging pilot results! But, is your #AHT really going down because of the AI? Or are there other factors driving this initial success? ?? We all know that pilot data is key to making the right long-term decision. But the integrity of that data will dictate whether you succeed—or fail. Is your vendor giving you objective data? Or are they giving it a little "spin" to get you to sign a long-term contract? ??♂? How are you measuring the objectivity of that data? ?? Let’s bring in our favorite sales guy, Uncle Rico, to explain a few ways to make sure you’re buying something that actually works — and not just another 24-piece set.

What's the most important phase when developing AI agents?

Salesforce Agentforce → Overrated or Underrated… PART 2 I still have so many questions for XtendOps AI expert, Giovanni Toschi What’s the best approach to implementing GenAI in your CX strategy? Is a single-platform approach easier/better than DIY? Is reinforcement learning from customer outcomes the new norm? Is it going to replace human feedback? Who owns the data? Who owns the data model? And does it play nice with others if I want to leverage tech from other providers? Also, Gio has some explaining to do regarding Sophie…

Ready to get your AI agent up and running? Let’s talk about the build phase! Or like I like to call it Discovery and configuration. First, we need to understand what we're dealing with. Discovery is where we gather all the goods SOPs, historical conversations, agent learning material, workflows you name it and what we see through incoming requests, figuring out which intent to automate first. Next is configuration: We design the architecture to fit exactly the needs we have identified connect it to your CRM and other critical systems, and then we configure the AI for specific data and intents. Finally, we test it out and get your sign off. So how can you use this to your advantage? By ensuring every step in the build phase is meticulously planned and executed. You set the stage for a seamless automation journey. Discovery helps you avoid costly mistakes by focusing on what matters most: high volume, high impact tasks that offer immediate ROI. In the configuration stage we fine tune the AI to speak your language, literally and we make sure it's connected to all your system. So that there are no gaps in services or knowledge. If you skip or rush through the steps, you might automate the wrong process. But when done right, the build phase ensures your AI. It’s not just functional, but a powerful tool that aligns perfectly with your business goals. After all, automation is the journey, not the destination. #xoai #technologio #build

September marked the 5-Year Anniversary of XtendOps When I was growing up, my family almost always spent New Years Eve/Day together.? My youngest brother, Mike McCreery (aka Mikey), would always ask the dinner table?the same question ‘what was your best memory from last year’?? That same brother helped me start XtendOps. So in the spirit of a new year for XtendOps, I thought I’d answer that same dinner table question. My best ‘memory’ has been working & sharing this entire journey?with my little bro, Mikey. Thanks for believing in me when no one else did,?for always having my back, for always taking on the worst problems and for?being the best person I know.? But, since it's been 5 years, I have a few other memories I want to share…?

Agentforce → Overrated or Underrated?? Salesforce Dreamforce felt different this year. The application of GenAI in the CX space is becoming a reality—and it's happening fast. And everyone is struggling to keep up. The marketing game is strong. Use cases are still in the early stages, and data is preliminary. Is AI already driving gains, or are there other factors at play? When I left the keynote, I was genuinely excited. I felt validated. One of the biggest names in tech is focused on delivering AI agents at scale—just like us ?? While validating, I did have some questions for our XtendOps in-house AI expert, Giovanni Toschi

Snoop Dog and Agentic AI have more in common than you think,?? It's Artificial Intelligence that acts with agency. You basically need three things: One, you need a large language model. Two, you need something called cognitive framework, basically a framework of decisions and approaches that the AI follows to reach the goal in this case, we refer to the REACT framework, which stands for reason and action and third, lyou need tools, things that the AI can use to achieve the goal. A very easy example is Web search. Let's say that you have a question. How old will Snoop Dogg be in 20 years from now? The traditional way is that you would go online. You would Google and you would ask for the answer, and you will get an answer like 52. But that's not what you really want to know. You want to know how old it will be in 20 years. That's what agentic AI can do for you to answer the question. First. I need to know the age. There is no doubt. Do I know it today? No. And so I need to take an action. So if you go online, it will in the on your behalf and it would Google. What is the age of the dog, which would lead to a result or an answer of 52 years old? Then it will reason on the result and conclude that well, that's not enough to be able to say our oldest dog will be in 20 years. So it will conclude that there is another action to take, which in this case, is to calculate the final answer and get to 72. Now, apply this to something that is very complicated, and you start to see why the has a massive potential. #technologio #agenticai #snoopdog

Learn what makes ChatGPT different from a simple chatbot in less than 60 seconds ?? Traditional Chatbots were based on rules they can be “if then” rules or they can be “intent based" rules So when you say “hi” then the chatbot always replies with "hello" and there is really no way in which you can customize this. Or you can use a more advanced prediction system, which detects the intent in this example, if the user is saying "hi" then we detect there is an intent of type “greeting” And so instead of just always repeating the same greeting: "hello" we could have some sort of additional variable or rule that can change the answer based on something else for example, the time of the day. So it could “good morning” if it's morning, or it could be “good evening” if it's evening, and that's what the intent does. The whole point of ChatGPT is that there's no strict rule and there is no intent when the user says something: Hello, hi, ciao, bonjour... it is processed statistically and an answer is generated by ChatGPT based on that calculation but without any predefined rule, and they can generate something that is way more akin to human conversation simply because it's based on statistical approach and not on rules. #technologio #xoai #chatbots #chatgpt