La capsula Informativa: frankly… Episode 58: How are you using AI?
The transcript below is AI-generated and may contain minor inaccuracies. Tune in to the episode audio to hear the full conversation!
Transcript
Rachel
Hello, welcome to frankly.
Dan
Welcome.
Rachel
We have a solo episode today and we’re going to get into AI, AI and communications and marketing and really just kind of even the whole world of AI and implementing it in like with good practices.
Dan
Yeah.
Rachel
Because I think that can be one of the hardest things for companies and not knowing where to start because. Is it feels overwhelming and it feels like there’s so many rules which there is the way to use AI in like a correct way or appropriate way in a business setting. But we feel that used correctly it can absolutely be an incredible tool that actually allows. You as a communicator, to spend more time being strategic and doing the things that are important for a human to do.
Dan
Right. And and beyond, even just so many rules, I think the other piece of it is just so many tools, right, It’s overwhelming.
Dan
Like there is there are 100 different AI solutions probably for any given tasks that we want to do as communicators.
Rachel
Yes.
Dan
So what? What goes into that research process? What do what to look for? Like what? What do we want to see in AI to make sure that it is something that can be a good fit?
Rachel
Yeah. Yeah. And how do you decide which one to use for what? Right. I mean, there’s ChatGPT is the one that most people know, but there’s Claude, there’s Gemini, there’s Copilot, there’s and. And these are coming out. They’re all kind of that, you know. The like, like the large language models like, it’s all very similar, but they do have strengths and weaknesses compared to one another in their versions or in how they like you interact with them. That makes them more beneficial for certain things over others. So. We’re going to dive into that a little bit. I am a part of a group at Franco that’s really doing a lot of AI research and trying to figure out how we implement it as an agency. What to really just like living in these tools to understand their capabalaties and just like experimenting across the board and how to use them, so I’ll share some examples kind of of what we’re doing and how. But it’s been a really fun. Just process. I feel like I blew my father in law’s mind yesterday talking to him about AI. He and he. Was like, he’s like, how do I get this? On my phone. He’s like this. I could use this to research. I’m like it goes even deeper than that. Right? Like, you know, from a competitive standpoint, he’s in automotive sales. How are you researching other companies competitors, right?
Dan
Yeah, yeah.
Rachel
It doesn’t. Half he was like at one point we had an entire department dedicated to doing that analysis. I was like and not today, not today.
Dan
Right, right. And even if it could just get you started to the point where I mean, obviously you gotta check everything and make sure, but if it can get you to the point where you have all the resources or have a great start on the resources that cuts so much time out of that.
Rachel
Of course. Yeah. Yeah, absolutely.
Rachel
In my own kind of playing around with this and research, I’ve used it in a couple different ways and I wanna start with just some general education on ways to use AI. I actually put a sticky note on my desktop that goes through the 6 core Generative AI usages, like things that I think I had to put it on my desktop to remind myself, I think the hardest thing with AI, is remembering that it is a tool that you have it at your disposal that you can use it right because we are so programmed. How many years have you done your job that you just do your job and you’re not thinking about? There’s a tool there that could really. Make it. Faster or easier or just bounce ideas off of right? It’s kind of like the it’s a personal brainstorm, you and I. So the six core generative AI uses: generation. So this is like the most commonly thought of right making stuff. Write me 3 social posts for this client.
Rachel
These things, here’s our, you know, core messaging, whatever. That’s. That’s how it’s mainly thought of to be used. The next thing is extraction, so pulling data out of a source and reformatting it. So this I think. For me is I used this we did an internal survey and I took those survey results and I threw it into AI and I said, what does this tell me what so like give me, pull out the biggest data pieces, pull out whatever it is that you think is important, that encompasses what. The responses were of this survey, so it was pulling out kind of that I had it reformatted to send. Summarization. Same thing, kind of like that same example. Summarize it for me, right? I said pull out the key points. I also said summarize this in, you know, 5 sentences that I can send to our leadership.
Dan
Yeah.
Rachel
The 4th is rewriting, so turning one kind of data into another. I think this is really beneficial if you are writing a piece and you can use it both ways right? So the one way I like to use it is if you’re writing something, it’s just too long. You’re writing A blog. It’s. 1500 words.
Dan
Yeah, it needs to be 500.
Rachel
Yeah, you’ve got a lot to say, but it’s we’ve really got to boil it down, put it into AI and tell it what it needs to be, how many characters, how many words. And let it work its magic and then go check its work and make sure that those core points.
Dan
Right. What’s important is actually still there.
Rachel
Absolutely, because the biggest thing about this is, you know, it only knows what we put into it. Which be careful with that, right? Don’t put any proprietary data into it. You know nothing that’s not public information. We do know that if you pay for the tools, it is a closed loop, which means it’s not learning with what you’re putting into it. So if you’re paying, you have a little bit more freedom. But I don’t think it’s necessary for you can use it for so many ways without it being proprietary data.
Dan
Right. Yeah. And if you’re just testing things out and trying out, you know, free trials of tools or things like that.
Rachel
Yes.
Dan
That’s where, yeah, we kind of alter cautious on that side.
Rachel
The other thing with rewriting that I think is interesting that you can do is if you have a client or you work for a company that does a lot of owned content. So blogs and then you have gated resources, right, your white papers, your ebooks, that all is a little bit different, right? Like White Paper is a little bit more. Actionable. How do I turn this information into actions I can use? If you have a blog and we talk about maximizing content, throw it into AI and say I need to write a white paper, I need 5 actionable tips that come out of this blog.
Dan
Yeah.
Rachel
What are some options? Yeah. And then refine it from there. Right? Feed it a little bit more information, make it get more specific, but that’s a really good and easy way to not duplicate your work, but get multiple content pieces out of it.
Dan
Right, right. So you start with an original piece. You start with, you know, all of this thought leadership and then you can really boil it down into what what people can take away, yeah.
Rachel
Yeah. Yeah, absolutely. Or if it’s a blog or white paper, you can, you know, put it the opposite. Or you can say I’m an executive turn this into a thought leadership LinkedIn article. Right. Get it down to 500 words.
Rachel
So there’s a lot of really awesome. I think when that like with that rewriting that is communicators is right at your fingertips classification. So organizing your data right, if you’ve.
Dan
Yeah.
Rachel
Got a big? Excel document full of all of this data that needs to be organized. Put it in there and you have to make sure you’re telling it what the output needs to be. I think that’s one of the most important things. When you put something in here, you have to say I want my, you know, I want the output of what I’m giving you to be an Excel document to be a table, to be a whatever it is you’re looking for. Get specific.
Dan
Yeah. Give me these 4 headers. Give me this you know total sum at the end, whatever it might be that you want to be able.
Rachel
Yes. To yeah. And last but not least is question answering that simple, right?
Dan
Yeah.
Rachel
If you have a question and and we’re seeing this even in Google Now, if you Google something, if you put a question in Google, the first thing is AI, right? It’s taking all. Yeah, right. But it’s taking all of the stuff on the Internet. All those pages and trying to get you the most responded accurate could be one thing or another.
Dan
Yeah, with mixed results, yes. Yeah, yeah. Right, right.
Rachel
We’ll say it 100 times check AIS work. Yes. Check it. Tell it to check itself. That’s actually something I’ve learned too. Yeah. Like, go back and check your work.
Dan
Yeah. Rule number one, yeah.
Rachel
Did you miss anything? Was this computed correctly? So answering questions and you can do that about data too, right? If it’s data or just general question about. I know we’ve used it for research purposes, so we have a lot of very niche B2B clients.
Rachel
That. You have to do a lot of digging or you have to ask and spend a lot of time that you know put it on the client to explain things to you or get you that information. So not everything’s always public information on the Internet still do that, but asking it to explain a process or whether it’s a manufacturing process or the history of something or whatever it might be. Ask a question.
Dan
Yeah, and and on the research side, I think one thing that seems really valuable is to be able to find data points that support a message that you have. So even if you got, you know, even if you’ve got your storyline down, but you need to say like you need to find a few stats on what’s going on in the industry, what’s going on in the world to support kind of this trend that you’re working on, you know what’s out there? But it’s just sometimes it’s hard to find those types of things.
Rachel
That’s a good point. Yes. Yeah. Yep.
Dan
Source. So if you can ask it to kind of find stats on a certain thing that can be also nice research tool there.
Rachel
Absolutely. It’s a great point. So those are your 6. Core uses. Now there’s more than that, right? It depends on what you’re doing and how you’re doing it, but I think there’s a few ways that I’ve used it recently that have really been helpful. And again, it’s not the final output. It has to be dug into.
Dan
Right.
Rachel
But. The first one is meeting notes and action items.
Dan
Yeah, yeah.
Rachel
So there’s a couple ways of doing this right? Like Zoom has a tool that will summarize or give you a transcription or anything along those lines that you could take and use on its own. But say you don’t have Otter.AI, you don’t have the Zoom AI feature you don’t have, right? Like you don’t need it to do. This piece and I’ve played around with this, you know, using that versus actually taking your meeting notes. So you’re gonna take meeting notes regardless. Most of us usually if you’re in a meeting. Taking those meeting notes and putting that into AI and saying please give me action items for us and. Client or company or vendor, whoever it might be based on these notes and it will do actually a pretty good job summarizing it. You can condense, you can take away right? It’s not going to be perfect.
Dan
Sure. Yeah.
Rachel
But versus you kind of scanning through everything, extracting what those are putting those and then following it by meeting note. It’s a really, I would say it can save you close to 10-15 minutes. Which?
Dan
Hey, if you’ve got, you know 5 to 10 calls in a day, that can be a huge time saver. That’s, you know, an extra hour to do things that are actually meaningful.
Rachel
Any bit you can take. Absolutely. Are actually meaningful? Yeah. And if you’re not sending your clients or whoever it might be. Go up. Action items following a call really good time to implement it and still be able to save yourself some time, right? Not add any extra time to you know what you’re already doing for them. But it’s a great value add that you are providing as a vendor that could you know, just the professionalism is very appreciated, I think.
Rachel
You know, just the professionalism is very appreciated, I think.
Dan
Sure, yeah. And and even just within your own team, like if there’s not a, you know, even if there’s not a client where they involved like. Being able to list those out and just be able to quickly and easily log who within the team is responsible for what, so there’s no confusion. There’s no, nothing’s lost in translation like that is a great take away from a meeting like that where you can just easily assign and get on your way to doing the work.
Rachel
Yeah. Totally. Yes. Yes. And then if you take it one step further, say you do have access to tools or automation or whatever it is. The ideal and what I’d love to get to is being able to take meeting transcriptions or meeting notes and then have it automatically send a Teams message or something telling people these are your assignments right? Like adding it to Microsoft to do whatever that might look like so you can take AI if you really get granular.
Dan
Yeah.
Rachel
Even further to help automate process. Processes that could really save you a good chunk of time in that kind of project management sphere, so you could even take your meeting notes from the last time and have it make an agenda right. Work backwards what is, you know, what would we still need to talk through? What did it say was outstanding that was going to be completed. Create an agenda for me to work through.
Dan
Right, yeah.
Rachel
Again, you have to edit it, but it’s. Not that first. General, you know, think about we said the first core use is generation, right? Use it for that versus you writing it all out. So that’s one I think really good way to use it. We’ve talked about a lot of data extraction and summarization. So think about your reporting processes. I think there is a huge. opportunity for being able to take if you export out of LinkedIn or Instagram like think of your social channels or MailChimp and take that and put all of that those numbers into AI and tell me what are my averages? What is this telling me about my this send or this social post or whatever it is?
Dan
Right, right. Yeah. So. So question about that for you, like when you’re looking at reporting technologies of you know that are existing that we already use that kind of connect with LinkedIn and kind of have that integration or that API already like.
Speaker
Yeah.
Rachel
Yeah. The. Yeah.
Dan
What are you seeing that AI can maybe add to that beyond what was in an existing reporting tool or aggregator?
Rachel
Great question. To me the value add is the analysis.
Dan
Rachel
So if you can put in your and yes, we have a great digital team that can do that. I’m sure if you’re listening, you do too. But again, think about the time it’s getting them one step further versus having to go all the. Way back so. In putting all of that data, you can see in your reporting tool and saying analyze this, we summarize it. What is it telling me?
Dan
Right.
Rachel
Did all of my posts that had sliders on LinkedIn, did those perform outperform video posts?
Dan
Yeah.
Rachel
You can see that by looking at the data, but it’s going to take you a minute to find the pattern versus A.
Dan
Yeah, and.
Rachel
I’ll find it quicker.
Dan
Yeah. And that that makes sense. So if you, you know using that example, it’s just format, right? Yeah. If you get the idea of what format is working best, then you can go back and look within that format to see what the where the content advantages were or what the message maybe was in some of them that even outperformed the average of the highest.
Rachel
Correct.
Dan
So that’s very cool, yeah.
Rachel
Exactly. I think that and that’s a great question because it’s not only saving time, but it’s also just. Smarter, right? Where again, that human eye is still so necessary. But if it gets you that one step ahead for you to go back and draw those conclusions yourself, that’s exactly what you.
Dan
Yeah. Want. Yeah. And I think that was what my question was rooted in. Like. Yes, you can automate things. Yes, you can use AI for some of these things, but you know. Is there a value above what already exists in? Maybe a non-AI driven tool? And I think that’s a great example of like where the actual just data aggregation might already exist, but it’s, you know, one step closer like you.
Rachel
Yeah. Agree.
Rachel
Another opportunity that I’ve used is. Recently someone came to me and said I need. Basically like I can’t even tell you what I need, right? But here’s what I’m trying to do. I want people to sit in the room. I want them to. I need to ask them questions that get me this information, right? But I don’t even really know how I want to organize the end information. Help me.
Dan OK.
Rachel
And I said OK. That’s fair. So I went to AI and I want to know in my experimentation, Claude is really good for long form content. So what ChatGPT is going to be a little bit shorter? I’ve used that a little bit more for.
Dan
Rachel
Just. Like asking questions that that like generation of like. Sure, whether it’s social posts or like another example that I think is interesting to get the ideas flowing is I’ve done advertising copy in there. So I’ve said to it like, you know, I’m this kind of company.
Dan
Yeah.
Rachel
And do these things or like, here’s an example of a really quick, you know, the information I want in the ad, but it has to be 30 characters. Yeah. Help me get there. And then I’ve said, like, be more witty or I’ve said like be more playful or be more serious. And you can really tweak it that way, which I think is like another fun way of using it.
Dan
Yeah.
Rachel
Chat GPT was pretty good for that.
Dan
See, that’s one of my favorite and and I am not as active of the user yet I’m interested, but I’ve just you know, like you said at the beginning, sometimes I’m not sure like it just doesn’t pop into my head all the time.
Rachel
Yes. Of course.
Dan
But one of my favorite things been able to use it for. Is that really like summarizing in a few words? So I love writing. I hate writing headlines, subheads like all of that stuff.
Rachel
Yes. It feels daunting, like how do I tell this story in this many words? Yes.
Dan
Right. How do I sum it up in 5 words, 6 words, whatever it might be, you know, for newsletter subjects, e-mail subjects, blog titles, whatever it might be. I’ve that is one place where I’ve really found value in it, because that’s the one thing that.
Rachel
Yeah.
Dan
Like you know, it’s not stressful, but it just takes me more time to write those 5 words than it does to write 5 paragraphs. And I love being able to say, give me 5, give me 10 options and maybe you don’t even pick one of them. Rachel
But it sparks an idea.
Dan
Yeah. It either sparks an idea or you can kind of mix and match between the options to really come up with something that tells.
Rachel
Yeah. The story well, I’ve even found if, like the creative, the initial. Well. You know, words that I want to put on a creative graphic and then the, you know, designs like, hey, Rachel, that’s just too long. Like within the space they’ve given me because I’m looking. I’m not looking at the dimensions necessarily and being like it can fit this many characters, but they are.
Dan
Yeah.
Rachel
Yeah, and there’s like. Hey, it’s not fitting on here. I know that headline and that subhead are good, but. I just need it shorter, right? That’s like the perfect time to use because I would sit there and play with it forever to get it right. So another great example. But going back to kind of that like what you know, I don’t know what I want my output to be, but I need this to be actionable and I need it right, like figure it out for me.
Dan
Exactly, yeah. Yeah.
Rachel
One thing that I think we all have to keep in mind when you do anything like this example is your prompts AI and what it gives you is only as good as what you put in. The prompts are gold. You have to be specific. The more specific you are. The better it can serve you, just keep that in mind when you are writing this. So I went in. I told it, you know who I was, what I was trying to do, all that good stuff. And I said, I want you to outline what this output would look like, right. Think of a strategy.
Dan
Yeah.
Rachel
What is the business strategy, you know, give me an outline based on this information that I’ve given you and who I am and what I’m trying to do. What is a business strategy look like for 2025? What are the core components that I need to have in it?
Dan
Yeah.
Rachel
It goes through it spits that out. I say. Now here’s even a little bit. Go one step deeper. This I want you to focus kind of in this area. And what I think Claude did well when it spit that out for me was it said at the bottom it actually gave me like, here’s some ideas of how you could adjust this
Dan
Oh, that’s interesting.
Rachel
So do you want more here or are your services more focused? Here.
Dan
Right. So it’s kind of prompting your prompts almost.
Rachel
Yeah. And then it actually told me when I said from there, like, OK, now I want you to put some discussion points together that I can walk through with my team that get me the answers to these things. Right. To fill this in, what questions do I need to be asking right. So I had it go back, then it actually told me, which I thought was really cool at the bottom. Here’s what the pre work that would need to be done like here’s the information that you need to provide your team with to be able to productively have these conversations and answer these questions which was not something I’d ever thought of right. Like I could look at and be like oh, you need this, this and this, but because it’s generating those, it knows what it’s based off of which goes back.
Rachel
To that top right like. I guess my point is like you can work fully backwards into something like I’ve.
Dan
Yeah.
Rachel
Said to me, I want this thing from you and.
Dan
Yeah.
Rachel
I said OK, I’m gonna. Go back like 5 steps.
Dan
Right.
Rachel
To get the best this thing. So then from there I kind of refined it a little. Bit. More and I was like, OK, but. This is really long. So I asked it if you had to summarize this in five like what are the five top questions that I should be asking that get me the most answers to? This.
Rachel
And let it spit that out. Yeah. And again, it gave me, like, the information that you would need to give in advance the pre work, all of that kind. That stuff. So I was able to provide three things to an ask for one thing, right. Yeah, the full outline. The long Question Form Questionnaire and the, you know, shorter one. Yeah, that is now, I did that, and I did tinker around for a while, kind of on purpose because I was exploring. But it was less than an hour?
Dan
Yeah, and. And if you think about the time it takes to. Not just think of those questions, but, you know, gather them all up, you know, understand what falls under what category. Get the format down and then boil it down for. Yeah. Exactly. And then boil it down into like, the really critical areas. You’re even if you spent some time tinkering in the tool, that’s probably a fraction of what you would have spent.
Rachel
I was saying not to mention. The format. Absolutely.
Dan
You know, getting all of the. Together this manually.
Rachel
And I was able to provide. I mean you asked earlier like what’s the value add. I was able to add value and saying like here’s what I think is good right like you kind of gave me creative freedom. Yeah. But here’s some options and here’s what I got to that. Like this is what the structure would look like if we answered these questions and did. This in the way that it thinks is best. This is what hat you would have as an outcome.
Dan
Yeah.
Rachel
Which I think is. That’s like the value add that I didn’t make that person have to figure that or or tell me that they didn’t have to spend the time doing that? Yeah. So across the board. And I learned something, right. Like I learned if I want to put a business strategy together, what does it look like? What should I be?
Dan
Right, and now you have some of those, now you have some of those questions for next time.
Rachel
Yeah.
Dan
I mean you know each business is obviously very different, but like the the core questions that drive strategy often are the same.
Rachel
Yes.
Dan
I think you said something interesting there with like the mix of creativity and efficiency.
Dan
Right and you know, prior to this, everything we did was manual. You know, you’re not going to automate everything in the world.
Rachel
No, and not right away.
Dan
But if you can find, like the exact balance or not the exact balance, but the kind of best balance for you of how much you can input from a creative standpoint, what you can automate on more of that administrative side and. Kind of create the best of both worlds between those two. That’s I mean, you’re going to add value in some amount. You’re going to add efficiency in some amount. You know whether it’s big or small, it adds up.
Rachel
Absolutely. And how do you measure that? That’s a whole other thing we’re trying to figure out. How do you decide whether your ROI and investment in that is?
Dan
Yeah, yeah, maybe another podcast. Yeah.
Rachel
Actually, you know the time saved is allowing you to bring in more dollars of strategy work or whatever it is that’s still TBD, right? That’s a little bit more down the road, but.
Dan
Yeah, you gotta have time to let it work and let you know. I guess a get a good process that everyone can go through and then B, you know, give it some time to breathe and gather data. So you can.
Rachel
Yeah. Totally. Look at that. Two more quick examples that I’ve used that I think are helpful, kind of different. So one is if you have the transcription or excuse me if you have the. Recording of a meeting right say to info gathering, meeting to be able to do a media pitch right. And so you’re talking to an SME.
Dan
Sure.
Rachel
You record the meeting, you can use AI to do transcripts, right? Give you a full transcription of the meeting. That’s. Great. I have then taken that transcription and dropped it into AI and said give me the key messages that came out of this. What are my statistics? What are my talking points? Give me like the top level messages. Depending on how the conversation flows, it’s not always getting it perfect right? Then I said it gave me, you know, like 30, I said give me put these now into 15, right. Yeah. And it still you you still very much have to go back maybe you you know listen to it again as you’re editing it.
Dan
Yeah, yeah. Let me get tenants.
Rachel
And how well, how good your transcription is as a whole other thing, right? Like, yeah, they’re a specific example of this where it didn’t work so well is our SME had a thick accent.
Dan
Yeah.
Rachel
The transcription then gets a little muddied, which then you know, so just be smart with how you’re using it. But another great opportunity for boiling things down. Taking a big transcription and and getting something out of it. So there’s just tons of opportunity that really. With each tool itself working through kind of what its ability to do, like Microsoft Copilot in your outlook right now, if you have the new outlook, yeah, not so great. Not gonna lie. That has a little work. I honestly, I asked it to.
Speaker
Yeah.
Rachel
Which I thought this was a great idea and should be able to do compare calendars like I need to meet with this person for 30 minutes and the next week tell me you know. Times that work.
Dan
Give me a time, yeah.
Rachel
How many times does your client like I need times to meet and you’re like, you know, manually going through comparing calendars again, not something it takes like that much time at the end of the day.
Rachel
You know, you gotta you. It it would be a lot easier if it.
Dan
Would save time, yeah.
Rachel
Could just look. At it. So I think there’s a lot of and and it they’re change. They’re always releasing updates. They’re always trying to get better at these things, so.
Dan
Yeah.
Rachel
All that to say. Would love to know if you have some really good generative AI usages.
Dan
Yeah. Let us know.
Rachel
Let us know. Send us a message on LinkedIn, comment on our posts. Leave a comment here. But I feel like there’s this episode could have like 7 follow-ups from here.
Dan
Ohh yeah and I’m sure it will, in time.
Yeah. And once we roll everything out at Franco fully and we’re working toward it like I I would love to do a follow up to this and kind of let you know where we’ve landed.
Ohh yeah and I’m sure it will time.
Rachel
Yeah. And once we roll everything out at Franco fully and we’re working toward it like I I would love to do a follow up to this and kind of let you know where we’ve landed and how we’re using it, but. Don’t be afraid of AI. It’s we know so much more about it even than a year ago. I think about a year ago, like I was terrified of it. Like, how am I gonna ever use this? What are its opportunities? What are its misses? It really has so many more opportunities if you use it in a good way. So let us know how you’re using it. Don’t be scared of it and have fun experimenting. Yeah, AI can really, really be an awesome tool, all right. That’s it for us today. We’ll be back in two weeks and.
Dan
Yeah.
Rachel
We’ll see you then.
Dan
Yeah.