The Digital CX Podcast

The Generative AI Primer for CSMs | Episode 055

June 04, 2024 Alex Turkovic Episode 55
The Generative AI Primer for CSMs | Episode 055
The Digital CX Podcast
More Info
The Digital CX Podcast
The Generative AI Primer for CSMs | Episode 055
Jun 04, 2024 Episode 55
Alex Turkovic

Send us a Text Message.

In this, my second solo episode - I wanted to spend some time trying to demystify Generative AI for CSMs and CS leaders alike. According to statistics, only 25% of CS workers utilize AI in the workplace on a regular basis, which I think is WAY too low - especially considering how stretched thin most CSMs really are.

So - in this episode, we focus on proper prompting, a few tools that exist out there and a plethora of use-cases for you to dig into:

0:00:00 - Introduction
0:03:55 - Topic introduction. Why GenAI for CSMs
0:08:42 - Why prompting is a fundamental skill to have
0:09:36 - Using the RISEN framework for prompting
0:11:31 - Taking care with proprietary and sensitive information when using Gen AI
0:19:58 - Why it’s a bad idea to just copy and send, without proofreading and personalizing
0:23:01 - Utilizing ChatGPTs memory feature to prevent having to copy/paste your prompts
0:23:55 - Teaching ChatGPT on my tone of voice
0:26:19 - Chaining prompts
0:27:55 - Integrating this into your daily workflow
0:30:21 - ChatGPT vs. Perplexity vs. Google
0:32:42 - Perplexity research use cases for CSMs
0:36:30 - The proliferation of new tools
0:37:19 - AgentCopilot & HeyGen create personalized video for your contacts at scale
0:39:45 - Ariglad analyzes support tickets to create and update knowledge base articles
0:40:48 - Malik automates the creation of decks using your data and insights
0:41:55 - Outro

One link discussed in the show is the DCS Tech Stack on the website: https://digitalcustomersuccess.com/tech-stack/

Enjoy! 

Support the Show.

+++++++++++++++++

Like/Subscribe/Review:
If you are getting value from the show, please follow/subscribe so that you don't miss an episode and consider leaving us a review.

Website:
For more information about the show or to get in touch, visit DigitalCustomerSuccess.com.

Buy Alex a Cup of Coffee:
This show runs exclusively on caffeine - and lots of it. If you like what we're, consider supporting our habit by buying us a cup of coffee: https://bmc.link/dcsp

Thank you for all of your support!

The Digital Customer Success Podcast is hosted by Alex Turkovic

Show Notes Transcript Chapter Markers

Send us a Text Message.

In this, my second solo episode - I wanted to spend some time trying to demystify Generative AI for CSMs and CS leaders alike. According to statistics, only 25% of CS workers utilize AI in the workplace on a regular basis, which I think is WAY too low - especially considering how stretched thin most CSMs really are.

So - in this episode, we focus on proper prompting, a few tools that exist out there and a plethora of use-cases for you to dig into:

0:00:00 - Introduction
0:03:55 - Topic introduction. Why GenAI for CSMs
0:08:42 - Why prompting is a fundamental skill to have
0:09:36 - Using the RISEN framework for prompting
0:11:31 - Taking care with proprietary and sensitive information when using Gen AI
0:19:58 - Why it’s a bad idea to just copy and send, without proofreading and personalizing
0:23:01 - Utilizing ChatGPTs memory feature to prevent having to copy/paste your prompts
0:23:55 - Teaching ChatGPT on my tone of voice
0:26:19 - Chaining prompts
0:27:55 - Integrating this into your daily workflow
0:30:21 - ChatGPT vs. Perplexity vs. Google
0:32:42 - Perplexity research use cases for CSMs
0:36:30 - The proliferation of new tools
0:37:19 - AgentCopilot & HeyGen create personalized video for your contacts at scale
0:39:45 - Ariglad analyzes support tickets to create and update knowledge base articles
0:40:48 - Malik automates the creation of decks using your data and insights
0:41:55 - Outro

One link discussed in the show is the DCS Tech Stack on the website: https://digitalcustomersuccess.com/tech-stack/

Enjoy! 

Support the Show.

+++++++++++++++++

Like/Subscribe/Review:
If you are getting value from the show, please follow/subscribe so that you don't miss an episode and consider leaving us a review.

Website:
For more information about the show or to get in touch, visit DigitalCustomerSuccess.com.

Buy Alex a Cup of Coffee:
This show runs exclusively on caffeine - and lots of it. If you like what we're, consider supporting our habit by buying us a cup of coffee: https://bmc.link/dcsp

Thank you for all of your support!

The Digital Customer Success Podcast is hosted by Alex Turkovic

Speaker 1:

Hey, it's episode 55, another solo episode all about AI for CSMs. Let's go Once again. Welcome to the Digital Customer Experience podcast with me, alex Turkovich, so glad you could join us here today and every week as we explore how digital can help enhance the customer and employee experience. Enhance the customer and employee experience. My goal is to share what my guests and I have learned over the years so that you can get the insights that you need to evolve your own digital programs. If you'd like more info, need to get in touch or sign up for the weekly companion newsletter that has additional articles and resources in it. Go to digitalcustomersuccesscom. For now, let's get started. Welcome to episode 55 of the Digital CX Podcast. It's so great to have you back and love having you back every week. I appreciate you tuning in and listening and giving me all your feedback and, yeah, I just couldn't do what I do without you tuning in, listening, liking, subscribing, downloading, all that kind of fun stuff.

Speaker 1:

I don't know about you, but we are in the doldrums of summer already here in Austin. It's been relatively rainy here, but I think that is now stopping, so the 100 degree temperatures are soon to come. And, yeah, no school. So the kids at home, and I'm sure a lot of you are in the same boat. So, um, good luck for today. Um, I've decided to do something essentially every fifth episode where, um, it'll be a solo episode. Uh, if you'll remember, on episode 50, I did a Q&A show that was really loved all your comments and kind words on the back of that show. Putting out a solo show is definitely a different muscle and, you know, a unique skill set, if you will, but it's also interesting because you know you're putting yourself out there. I remember, I think a year ago or so, daphne Costa Lopez put out some solo shows and she expressed on those episodes how it was a little bit nerve wracking and I got to agree with her. That said, I got stuff to say. You know we've had a lot of incredible guests on the show. We will continue to have a lot of incredible guests on the show, but what you can expect from me every fifth episode is a solo show on things that are on my mind.

Speaker 1:

Today, specifically, we're going to be talking about generative AI, but specifically for CSMs. What I want you to walk away from this show with is some ideas and some things that you can go implement pretty much immediately to increase the usage of generative AI, either for yourself or within your teams. So we're going to be digging into that today. One quick kind of housekeeping note I will be speaking this Thursday so, by the time this shows out, in two days at MADX Scale and CS Summit. That is a virtual summit that you can go sign up for. I'll put the link down in the description. It's some amazing, amazing speakers there. So I'll be doing a session on, essentially, the combination of human and digital and how the best digital emotions have humans involved. So I think that's all the housekeeping for right now. I look forward to hanging with you over the next 20 or 30 minutes as we dig in deep on some Gen AI stuff. So we're going to start this conversation calling it a conversation because I hope you interact with me a little bit on this after the show's live with just an overall why why am I talking about this?

Speaker 1:

But then also, you know, like the why of Gen AI in CS, and I think that's fairly obvious. But I did run across some recent statistics that were kind of eye-opening in that and I'm going to paraphrase this a little bit. But it basically said that one-third of SaaS software workers. So workers in the SaaS space use generative AI on a regular basis. Okay, kind of low, low if you ask me. But then it went on to say that only 25% of customer success workers within SaaS use AI tools on a regular basis, and for me that was a huge eye-opener, because you think about CSMs and what CSMs have been dealing with over the last few years and headcount reductions and those kinds of things, and in general, you know, we talk to CSMs, they're just swamped so much on their plates and so many plates spinning at the same time that in my view, any tool that can help alleviate some of that would be a tool that a CSM and CSM leaders would jump all over to implement and to pull in and to really drive throughout the organization, drive throughout the organization.

Speaker 1:

But I think there's like this air of you know myth around things like chat, gpt, and I don't really know whether you know we're afraid of it or whether we just don't know enough about it, whether we can't even come up for air to really learn about it a little bit. And so for this episode, that is my why. My why is to try to get that 25% north of 25%. You know we're all goals driven. I'm not going to give myself a goal and whatever it's hard to measure anyway. I don't even know where they got those stats from, but the point is I want to. I really want to help dispel some of the myths that are out there and also just to give you tangible ways of starting to use these tools that we all have in front of us now to help your workflow and help your day-to-day become better.

Speaker 1:

I think by now, the talk track of AI is here to replace me and digital customer success is going to replace my role. That's kind of run its course, and I think more and more people are getting hip to the fact that A these generative AI tools are here to help Like. They're here to help make you more efficient, and that's what most of the tools that are built on top of these, you know algorithms are there for a specific reason to help workflows and to help people be more efficient and to help CSMs get out of the weeds and get out of, you know, this place where they've been struggling from a day-to-day perspective and into a place where they're consistently driving customer outcomes and customer value and having those valuable conversations right. So that's really my why I want to increase that percentage. I want to get things north of where they are today so that these tools can help you to succeed in driving customer outcomes. Look, the other factor here as well is, as artificial intelligence becomes more and more part of our daily and regular lives, employers are going to start really honing in on people who know what they're doing in this realm and know how to use these tools and know how to implement these tools, and so I think, more and more, you're going to start seeing that as a skill set that is a value add to an employer and, given the state of the job market today, I think that is something to pay attention to for sure. So where we're going to start with this is in prompting, and that may seem like a weird place to start when it comes to a show about AI in CSM or in CS.

Speaker 1:

I think that one of the fundamental places to start getting into using AI on a regular basis is just using it on a regular basis, right, and a huge part of using it is being able to prompt correctly. Now you hear a lot of things out there about, you know, prompting and prompt engineering and all of that, and that is you know. It is definitely a skill and one that needs to be practiced and one that is important to understand. So what we're going to do is give you an overview of a framework that I've been using lately, and I've actually written about it a couple times on LinkedIn and also the website, but it's a framework called Risen R-I-S-E-N. It is pretty well known out there. It's taught quite a bit. You can find a lot of content on Risen if you search online for it. But Risen is for me and, and out of the various frameworks that I've tried, for me, tends to be one that leads to the most consistent results when it comes to the outputs.

Speaker 1:

So there's so many people that go into chat GPT and just say things, like you know, write me an article about the Concord or write me an article about so-and-so right, and, of course, with that kind of prompt, it's more of. At that point, chatgpt becomes more of a novelty than anything, because, yes, it can do it. Look, yay, is the output good? Probably not, and why? Because you haven't given it any kind of specific parameters in which to operate, in which to write this article, um, and, and you really need to in order to get out of it what you need to be useful. So, in talking about prompting, this is a skill that isn't just chat GPT, but it is a skill that can be used pretty much across any type of generative AI platform where you can actually interact with the model itself, and so this is highly transferable skill to other places, not just chat GPT. We're just using that as an example, because I guess it's the most famous one, I don't know.

Speaker 1:

So one thing that I do want to say before we go through RISN is an area of concern that has hit most of us in SaaSass, but in other sectors as well, and that is essentially, you know, something that I've encountered with my employer is is the um, is the the release of proprietary information into something like a chat gpt or, you know, bing search and all those kinds of things? Because essentially, what you don't want to do is to use your personal ChatGPT subscription if you have one, or whatnot and put in a bunch of proprietary information about the company, but also sensitive information about your customer. That's what you don't want to do. Most companies now are starting to get around to having an internal facing version of either chat GPT or Microsoft solution or whatever that, that that avoids the model being trained on what you put into it right. And so I say all of this with the caveat of make sure that, if you're implementing this in your work environment, that there are some clear AI policies out there, that you're adhering to them, but also that you're just being careful. It's much like in the US. You wouldn't want to release your social security number out on the internet, even though probably everybody has it already anyway, but you wouldn't go openly kind of posting it somewhere. And it's much the same way you wouldn't want to post your best customers, most sensitive information out into this model that's being trained on it, because then other people will have access to it too. So that's my quick caveat.

Speaker 1:

Risen R-I-S-E-N. Okay, we're going to go through each one of these. If you're watching this on YouTube, I'm going to put up an image of it as well so that you have it for reference there. But essentially, r stands for role. This is the role that you would like ChatGPT to assume in giving you a response. Some example of this would be play the role of a CSM. Play the role of an expert in hydrodynamics. Play the role of a Marine biologist play the role of a master, um, uh, chess player, whatever it is Right, uh, and and and really, this would be basically like if you were asking this question of somebody who would you most likely want to ask the question of who has the expertise that um that you are after? And by uh, defining what role it is, you're basically telling chat GPT to kind of ignore all this stuff over here and here and focus in on this area of expertise and you'll get a much more pointed response. So R is role and we go to I.

Speaker 1:

I stands for instruction. This is the basic instruction of what it is you want to have returned to you. Okay, so it can be. Write me an article about photography. Write me a synopsis of the movie Gone with the Wind, you know. Write me an email based on this call transcript that I've just given you. You can also feed it certain documents. So you can feed it a you know a contract, for instance. Or you can feed it a Word document. You can feed it a PDF. You can feed it an image, even, and say you know, write me a summary of this so that I don't have to read the whole dang thing. So, whatever it is the instruction is write me something about this. Okay, you can also ask it to create images for you, right?

Speaker 1:

Image generation is a whole other thing that it allows you to do. I I've been using this a lot for like PowerPoints. If I want an image of certain something and I can't find it on Google or something, then I'll ask ChatGPT to create an image for me that depicts a certain thing. You have to be pretty specific about it, and image generation is notoriously horrible at creating words on the images. But just know that is something that you can do. So we've got an R I for instruction.

Speaker 1:

Next is S for steps. This is literally the steps that you want it to output to you. So, for instance, if I'm asking it for some kind of overview, I might say you know, include two short introductory paragraphs, followed by a bullet list of specific instructions to complete this task, followed by a conclusion paragraph, and what you will get as a result is two opening paragraphs, a bulleted list of whatever it is you're asking, and a paragraph at the end. You're being very specific and very pointed as to the steps that you want it to take in getting you those specific results. So that's R, that's I. That's S. Now we're on to E, which is expectation. This is setting the expectation of what success looks like in the output. So something like it should be a very thorough investigation or it should be easy to understand. You are basically setting the expectation for how you want the output to be and what marks it should hit. Some examples of what you could put in there is, you know, again, asking it to be very thorough or very easy to understand or highly technical or fit for a certain persona, you know, if you want it to make sure that a third grader understands it, then put those things in that expectation section.

Speaker 1:

Lastly, we get to narrowing. Narrowing is where you're going to close in the walls a little bit more and get the model to think about just specific things and ignore other things. We kind of started doing that with role right when we talked about role, kind of narrowing the scope and forgetting what's left and right. With narrowing, the last step in RISN, you're going to narrow that down a little bit more. So, for instance, you're going to tell it what language to use. If you want it to use Spanish or English or something like that which ChatDB 4.0 can now do, you're going to tell it what language to use. If you want it to use Spanish or English or something like that, which Chatuby 4.0 can now do, you're going to want to specify what kind of tone of voice it should use. Should it use natural sounding language, professional sounding language? You know what kind of I guess demeanor do you want it to have?

Speaker 1:

One incredibly important thing that you're going to want to include in this section is that you basically want to tell it not to lie to you. If you've been around ChatGPT for long enough, you'll know that sometimes the stuff it gives you is purely made up, like completely not factual made up stuff, and you know there's lots of thoughts about that. I'm not going to go into those. But what you want to do here is specifically tell it not to make stuff up, not to embellish, and you want it to provide you with accurate responses. And so you might say something like do not embellish or make up any information that you cannot verify for accuracy. Putting some kind of language in there like that will then have it not make stuff up, which is super important, and you know, on that note, I think that a lot of folks think that, um, you can just ask chat GPT to write you an email, for instance, and copy and paste that email and pop it in and send it right off. Right? Um, there's any number of reasons why you don't want to do that. Partially what I just mentioned about it just kind of making stuff up. But also, you know you're going to want the output to sound like you. You're going to want the output to be something that would sound like it came from you, and in order to do that, you still have to go through, read it, edit it, modify it a little bit to make the whole thing seem like it does come from you, and so it reduces the level of effort that you have to go in and make those modifications after the fact. So that's RISN Role Inst steps, expectation, follows, subscribes, comments, all of that kind of stuff. So if you get value out of the content, you listen regularly and you want to help others to discover the content as well, please go ahead and follow the show, leave a comment, leave a review. Anything that you want to do there really helps us to grow organically as a show. And while you're at it, go sign up for the companion newsletter that goes out every week at digitalcustomersuccesscom. Now back to the show.

Speaker 1:

Now you can get incredibly specific with this framework, and what a lot of people do is they'll write a really detailed and specific prompt once, keep it in a document somewhere, note where some of the differences need to be and where you know where you need to modify the prompt when you go to use it for this customer or this customer or whatnot, and they'll essentially write a very detailed prompt once and then reuse it over and over and over and over again. So you start to get some economies of scale there. But the more specific you can be in Risen, the better. Now, this doesn't mean that you have to use this type of framework every single time. If you're just using ChatGPT for like a quick resource or you need to get a little bit of information together about something, you can just basically tell it to do it and, you know, kind of have a conversation with it. But if you're looking for output, that's, you know a specific piece of content, you know exactly how it's got to be.

Speaker 1:

Risen Framework is a fantastic way to ensure that what you're going to get out of it is of high quality. And this leads me to a little feature that a lot of these chatbots now have, which is essentially memory. So, again, we use chat GPT. For example, the newest model 4.0 actually has a feature in there where you can train it and you can have it. Remember certain things about you and the way that you like to have your information presented or facts about you, kind of like saving them to disk by specifically telling it, hey, remember this or, you know, commit this certain thing to memory. Those are literally things that you can tell it to do and it will confirm that. Hey, I've committed this to teach ChatGPT to sound more like me and to imitate my writing style. Right?

Speaker 1:

So not long ago, I asked ChatGPT to go analyze a few of my articles that I had written, and so I gave it, you know, the links to those articles and crawled those articles and I asked it to act as if they were like a linguist or something like that. I forget exactly what it was. It was like a linguist or a vocal coach, I don't. I don't forget what it was, but what it came back with is a very detailed analysis of my tone of voice, my way of writing, my communication style and all those kinds of things. Right? So great, have a very detailed thing. Awesome, appreciate that insight into how I work.

Speaker 1:

Then what I did is I asked it a follow-up question, or asked it a follow-up task, which was to say, okay, now turn this into a prompt that I can use in the future to ensure that my outputs sound like me. And so what it gave me was a much more condensed version of that. That was essentially a prompt for my tone of voice. Now, not too long ago, what I would have done is I would have copied and pasted that and put it in a document and then, anytime I needed a piece of content, I would grab that you know bit of prompt and put it in there and, you know, have it prompt in my tone of voice. But with these memory features, what I did was I asked it to commit this to memory, and now, whenever I ask ChachiBT to create a piece of content in my tone of voice, it remembers that it has it in its memory and I don't need to go and copy and paste that whole prompt in there, and it just eliminates that additional step. Right, so you can get crafty with this and really start to. The more you work with it and the more things you commit to memory like this, you can have it become a much better assistant over time, much like if you were to hire an actual assistant and they would learn your ways of working over time. This this is this is very, very similar.

Speaker 1:

The other thing I want to point out, too, is that my example of teaching it my tone of voice is a very it utilized a very common practice of chaining prompts. So you know, I I first asked it to create an analysis of my tone of voice and then, after that, I asked it to then create a prompt for me out of that. And so you know that's a prime example of how you might, you know, chain things to where you're not asking this tool to. You're not giving this tool a massive set of instructions for what you want. You're kind of chaining things and improving things as they go or making things more like what your end goal is over time. So there's certain instances where you might, you know, use a big prompt if it's one piece of content, but there's other instances where you would want to chain some things to slowly morph the output into. You know what you're actually after.

Speaker 1:

So Risen is great. It gets a lot of great out. The outputs from Risen tend to be quite good, and it is something that you can use in various different platforms. Again, you might not use it all the time, but when you want a thorough piece of content that sounds like you and looks you know is is like the Pareto principle of things, then you're good to go, you're you're gonna want to review these things, and so. So what I want to get into now is is this notion of starting to adopt these things into your daily workflow, because it may seem very intimidating to say all of a sudden okay, now I'm gonna use ChatGPT for everything and I'm gonna put everything through ChatGPT, and that's probably not the way to go. The way to go would be to slowly start incorporating using it into your daily activities.

Speaker 1:

One great place to start is actually having it proofread your stuff for you. Sometimes you may just want to. You know you've taken some frantic notes and you've thrown together an email. You haven't read it at all. You can throw that stuff into chat GPT and ask it to polish it up a little bit for you. It's a prime use case for it and it's a great way to start using the tool Now. I think I gave an example a little bit earlier of how you might feed it like a call transcript, ask ChatGPT to create a summary for you and yeah, I get, there's other tools that do that for you. But if you don't have one of those tools, for instance, you might feed it that call transcript, ask it to create a summary email for you.

Speaker 1:

And what would have taken you 15, 20 minutes to throw together by you know, thinking back on the meeting, looking at your notes, looking at the transcript, making sure you got all the action items and all those kinds of things would have taken maybe a half an hour to put together. All of a sudden, chatgpt has done that for you in 20, 30 seconds. Copy and paste that. Then you can spend five minutes polishing it, making sure it's in the right format, it has everything that you want into it, inserting your tone of voice a little bit and sending that off. And so you've turned this maybe 15, 20, 30 minute activity into a five minute activity. And, on the aggregate, the more you pile those kinds of workflow improvements on each other, all of a sudden you're going to see some drastic improvements in your workplace efficiency. So again, you're not going to just grab everything that comes out of it. But what you are going to grab out of there is your 80% of the Pareto, where the majority is done for you. You just need to polish it and send it off and make sure it's accurate as well, make sure it hasn't lied to you. So when you think about chat GPT right, it is this assistant type, vibe. It is an assistant. You can treat it like a personal assistant. Write me this, compile this, summarize this, all those kinds of things, and that's fantastic.

Speaker 1:

One of the kind of downsides of ChatGPT is that it is, while it can search the internet kind of, it is somewhat limited in that capability, and so a tool that I have been using a lot kind of on the side of ChatGPT is a tool called Perplexity, which is essentially an AI search engine and, in fact, a lot of people are calling it an answer engine. So you think about Google, right, when you go search for something in Google, we are very used to the odd results that we get, and they probably don't seem odd because we're used to it. But you know, you might get an excerpt from an article, you might get a couple of links to some videos and you get some links to more sites. Most of those are paid, some may not be relevant at all, and then you get your organic links, but again it's all like links to stuff. Now I know Google is slowly shifting and they're just kind of playing with stuff shifting and they're just kind of playing with stuff. But for me anyway, the experience of using Google and Googling things has been just not very pleasant at all, and so I've really enjoyed using this tool called Perplexity, which allows you to, you know, has a lot of the same. It's basically a browser and a search engine, so you would treat it much like you do Google. But what it does is it compiles a lot of stuff for you into a generated answer and a very you know, very good and detailed answer. But it also provides sources and provides you with you know where it got the information from, so that you can then go read a source article, for instance, and so what it becomes is a fantastic research tool.

Speaker 1:

And where this comes in very handy as a CSM is a couple of different places. Let's say you're taking on a new account and you don't know anything about the company. Go to Perplexity, ask it to do an analysis of the company's recent earnings reports, for instance, especially if it's a publicly traded company, that stuff is out there so it can provide you with a summary of the earnings reports and the good. You know, the good and the bad. You can ask it to do you know a history of the company, for instance, and get some of that stuff out of there so you can start to do some pretty deep research in a very, very short amount of time. And that includes people as well. So, like, if there's kind of a well-known, if there's a well-known CEO, for instance, you can get to learn a little bit more about them. If you have a new champion on account, you might be able to do a little bit of research on the champion using Perplexity.

Speaker 1:

So it's this fantastic research engine that does use, you know, real-time web search as part of its results. So, as a CSM, having that duality of chat GPT and perplexity, where one is your kind of general assistant and the other is your answer engine, those two combined are really powerful tools to utilize and where chat GPTT kind of gave that warning of you know, look, you know, don't feed it proprietary information. Perplexity is a different animal altogether because in general you're using it as you would a search engine by trying to get information together about what it is you're going to talk about. Another great example would be industry research. If you're supporting an industry that's new to you, that you know nothing about, you can go really deep on that industry in a very short amount of time. So my recommendation would be go look at, obviously, use ChatGPT or whatever tool you have available to you, either in the Microsoft or the Google ecosystem as well, and then, you know, go check out Perplexity, because, honestly, I can't really remember the last time I truly Googled something. I've been going to Perplexity for just about all of that stuff. So we've covered a lot right, and my hope is that I've demystified some of that enough for you to where, if you're feeling hesitant about digging in, that you have some avenues now to go dig into that stuff.

Speaker 1:

I think, as a leader, these are the kinds of things that you might want to bring to the team. Heck, play this episode for your team and let me know how that goes, because once I put these episodes out, I don't know what happens. I've heard anecdotally some people use it for team meetings and things like that, which is super cool, but you just don't know. Anyway, the point is, you know, if you know of somebody who is really struggling with getting into Gen AI, hopefully some of these concepts can help them to at least broach the subject and start to get into it a little bit more. It's also a lot of fun, like you can have a ton of fun with generative AI. There are, you know, platforms out there that'll create songs for you so you can create your, you know, your team's theme song, for instance, and all those kinds of things. Like there's some super fun stuff out there.

Speaker 1:

I do want to point to three you know. So, okay, sidebar, there are such cool things happening below the surface in Gen AI, and one of the things that is a direct result is that the cost of development is starting to plummet, and so we're starting to see tools come out of the woodwork here and there. Now, the examples that I'm going to give you may not fall into that category, but the point is that what we're starting to see now is a lot of platforms and a lot of use case specific platforms starting to creep up out of the woodwork that are using these models in really, really cool ways. So I want to give you an example of a few that I've run across in the last few months that are doing some cool things. One of them is called Agent Copilot, and what Agent Copilot does is it's essentially AI-generated videos. Ai-generated videos, but you train it. Let's say I'm a CSM and I have, you know, 40, 30, 40 companies in my portfolio and maybe 100 points of contact. What I can do with a platform like Agent Copilot HeyGen is another one that does something similar is I can train it on my likeness and my voice, so I can give it some video to analyze and give it some audio to analyze, and it will essentially create you as a AI persona, right? I can then create a script. I can give it a script for a two minute, um, uh, let's say, for example, a product update video that I want to send to all my customers on the latest of the product, what's happening in the product. Um, then, through my crm integration, so my integration into salesforce or hubspot or whatnot I can feed it that customer contact data so first name, last name, email address, all that kind of stuff and what it will do is it will send a personalized, artificial, intelligent generated video to all of my contacts at the same time. So you've got a little bit of upfront work creating the script and integrating the contact list and those kinds of things, but then with the click of a button you can get a video that is personalized to each individual contact out to everybody on your list, you know, and it works like you know, like tokens. Basically, you say in your script, you say hey, first name or whatever, and then in the video it'll be like hey John and hey Mark and hey Mike and hey Julie. So it'll create a version for every one of your contacts. Super, super cool, very early on this type of technology, I think we're going to start seeing more and more of this to where you, as the csm, can make contact with a lot of people in very cool, innovative and engaging ways with very small amount of effort.

Speaker 1:

Another um, another platform that is very cool and much more focused towards our friends and support, which is this? This platform called AriGlad R-I-A-G-L-A-D, which what it does is it analyzes support tickets and the language and all that stuff in support tickets and automatically creates and updates knowledge-based articles. So if you're a knowledge-based manager, it's a cool tool to look into because it can dramatically decrease the level of effort it takes not just to create knowledge-based articles but to go through and update knowledge-based articles. If you've ever managed a knowledge base before and whatnot, you know what a huge lift it is to even make the smallest changes in a bunch of different articles. So AriGlad is really cool and a great use of GenAI.

Speaker 1:

I think we're all familiar with Matic M-A-T-I-K. They are an awesome deck creation tool where, essentially, you can feed it your customer data and your insights and have it create things like QBR decks automagically. You know, specialized implementation, kickoff decks and those kinds of things. So very powerful, because we all know, you know, creating PowerPoints takes forever sometimes. And look, this is just scratching the surface on some of these platforms platforms I do have a pretty extensive list of platforms like this and other categories available on the website. I think you go to digitalcustomersuccesscom slash resources and I think that's it. No tech stack go to the tech stack portion of the webpage and you'll see a whole list of, you know, cool, innovative tools that have, in some cases, very niche use cases that you might be able to utilize. So, look, I've meant for this show to be a primer and an inroads into using generative AI in your everyday work stream.

Speaker 1:

I hope that it has answered more questions than it has created. But in case it has created questions, please feel free to send me an email at alexatdigitalcustomersuccesscom and I'd be happy to chat more with you about it If there are some critical things that you feel like I have left off. Equally, send me an email If you're watching this on YouTube. Leave some comments down below and tell me where I goofed up and tell me all the things that I got wrong in this short segment. But again, I appreciate you engaging with me on the back half of these episodes, love hearing from you guys and again, I'll be doing more episodes, more solo episodes every fifth episode. So come back in a month and a week for the next iteration. I'm not sure what we'll talk about, but if you have any suggestions for that one, please do let me know. For now, I really hope that you've enjoyed this and we'll talk to you soon.

Speaker 1:

Thanks so much for joining and we'll see you later. Thank you for joining me for this episode of the Digital CX Podcast. If you like what we're doing, consider leaving us a review on your podcast platform of choice. If you like what we're doing, consider leaving us a review on your podcast platform of choice. If you're watching on YouTube, leave a comment down below. It really helps us to grow and provide value to a broader audience and get more information about the show and some of the other things that we're doing at digitalcustomersuccesscom. I'm Alex Tergovich. Thanks so much for listening. I'll talk to you next week.

Generative AI for Customer Success Managers
Effective Prompts for AI Interaction
Utilizing Risen Framework for ChatGPT
Utilizing Generative AI in Customer Success