Beyond dashboards: How generative AI is shaping the future of analytics for internal communications
SWOOP Analytics
APAC | Viva Engage Festival 2024
The emergence of generative AI is revolutionising work around the world, and no industry is left untouched. Analytics is one of those areas that have a tremendous about of opportunity to deliver even more actionable insights to a far greater amount of people than previously possible.
In this session Cai Kjaer, CEO of SWOOP Analytics, will paint a picture of where analytics in the field of digital workplace and employee engagement is heading and show you SWOOP’s current explorations in this area.
-
Our next speaker, which is Cai Kjaer, our SWOOP Analytics CEO and co-founder. Cai, do you perhaps want to take control and share a little bit about why internal communicators should be much more intrigued about generative AI and how we can help? Yes. So I think there's a couple of things that we've noticed that are happening, which I think are terribly exciting.
So, do I have to ask the question? Can you see my screen? Can someone just not yet? Does it come up? Yes. It's good. Yes.
Excellent. So it's, of course, what's happening in the world around us at the moment. And I think there's particularly a few things that I think means that we have amazing opportunities as internal communicators.
So the three things that seem to be converging in our time, which I think is incredibly exciting. After we had the pandemic that I think, fortunately, we've mostly forgotten about, if you know what I mean. But what it triggered was actually a huge increase in the rollout of tools to help communication companies.
That's one thing. A massive increase in the use of Teams and Engage across the world. Number two was that the concept of digital workplaces have now also seemed to mature a lot more.
And the combination of tools on these work in a much more integrated way than they did maybe 10 years ago. And then, of course, the last thing that's happened on top of this is generative AI. And it's interesting because it can generate things.
So if you're in internal communications, this also has a huge amount of opportunities. So I think this is coming up. When you bring these three things together, this is an unprecedented opportunity that in the corporate communications practitioners, we really need to have a serious look at.
And I just want to get a quick poll in around how you guys are using generative AI today, which will be useful for the rest of this conversation. So I've got a poll, actually two polls coming up from you. The first one is how you are using generative AI right now to do your work.
So you do it if you have to do your day job. So four options. Never ever, you don't want to have, you never care about it.
I'm experimenting. I've tried it, but not interested anymore. Or I use it every day.
And it's not every day for social things. It's for work. So if you can put your answers in.
If the poll doesn't come up for some reason, then just follow the chat message. And we have an alternative way of voting there. All right, let's see.
The answers are coming in. So what we can see is that there's a few, it's the no, never. You guys that have looked at it and said, no, it's not authentic or whatever.
I don't want to use it. And a lot of people are experimenting. Interestingly, no one has tried but given up.
And then nearly half of you are saying, we use it. We use it every day. So there's a nice, there's a kind of a nice split in here.
So for those of you, the next question is really for those of you who are experimenting. And for those of you who are using it every day. Let's go to the next poll and find out what you are actually using it for.
So the next question is really, the next poll is more of a work trial. And what I'm looking for is for you to give us some examples. One word, two words, three words of what you use it for.
For those of you that are experimenting or using it every day. I can see there's a lot of things coming in the chat. So a lot of questions, a lot of different things in here.
Get past the blank page paralysis. Yes, don't we all know about that one? Even though it must be, there's also now a co-pilot blank page paralysis, right? What am I going to type in here? Yeah, too many things to list. So many things you can do.
All right. Thank you for all of those. If we have a look at those, let's have a look at those entries.
And you can look at the chat as we sort of review this. I want to try and plot these against this plan, do, check, act model, which is really an old model from the 1950s, which is really something you can apply to activities that sort of happened to have a certain cycle to a lot of useful for business improvement. And if we look at these things, we can go like, so are we using generative AI for planning things? Are we using it for doing things? So that's more on the main part, I feel, writing content, preparing content, getting ideas for content.
Are we doing it for checking whether what we did actually worked? And are we doing it to help us improve so we can do things better next time? And I think as we look through this, derivative content, first draft, review comes, gives me ideas, researching and writing, ideas, copywriting, survey analysis. I think the way I see that, it seems like a lot of this is falling in. We'll probably put that in the doing part, if you agree with that.
That's from what, as far as I can see. The interesting thing, though, is here's one of the, I guess, conundrums that we are, I guess, struggling with as communicators. Because a lot of the focus, and that's certainly when I've talked to people, a lot of the focus sit exactly in that bottom right-hand corner.
It helps us produce things. The challenge we have is that there's huge opportunities for generative AI to also be used on the other half of the plan-do-check-act model, especially on the checking and the acting bit. Now, why does that become so important? Well, it becomes really important for two reasons.
One is that because generative AI makes it so easy to produce content, there's a tendency that we can now produce even more. Now, the flip side of that is that our employees, their attention span, their ability to absorb it has not changed. So, we've got to be really careful with all things we're sending out that it is actually working.
And I think generative AI actually has a huge opportunity on the analytic side to help us be much better at finding out what works and drawing on that so we can do our job even better. And I want to show you an example of that. I'm just going to change windows.
So, share. Let me share over here, share window. And then I'm going to bring this dashboard up.
So, I'm going to pretend that I am the comms partner for a guy called Ben. And Ben is a senior leader in school analytics. And it's my job as his comms partner to help Ben shine even more.
So, what I'm going to do, I'm going to find out and say, I've worked with Ben over the last three months to really help him prepare his really engaged activities. He has a desire to sort of have a more social side of him to start with and start having sort of more fun at work, if you like, and engages his team more. So, what I'm going to be doing, I'm going to go into threads and then I'm going to find our friend Benjamin.
And I'm basically going to find out all the things that Ben has been posting on Viva Engage in the last three months. Now, if you think about this from a typical dashboard perspective, these are the kind of things, you know, if you're collecting data on the use of Viva Engage or any of the other digital comms channels, you have a sense of these typical things and you'd be interested in finding out, for instance, I'm just looking at how many times Ben has posted and he's followed my advice to do it regularly, you know, not just once a month or something like that. He's posting quite regularly, that's really good.
I might sit with Ben as well and I'll say, well, you know, did we hit the right areas of the business when you were engaging? Ben, he works in development and operations and he got most of the engagement happening there. So, that's really good. This is, I guess, what we can say.
These are standard dashboards. You know, it's the place where you go in and you sort of have to know what you're looking for and you sort of have to understand the data to be able to really understand it. But it's not very dynamic as such, is it? So, I want to show you something instead about how we can use generative AI to actually help me as the comms partner engage with Ben.
So, I'm going to click on content over here instead and we're going to find the conversations from Ben that were the most engaging. And they come up here and instead of me going like, oh, like, what did you actually write about? You know, I gave him some ideas. Why don't we just click the little sparkler button, which is our Dr. Spook that Emily mentioned before.
We're going to click on this one and we're basically going to get our Dr. Spook, our intelligent coach, to help us basically summarise what is it Ben has written about and what is it that worked really well? And you can see here, instead of me thinking about it, reading, analysing, he'll just tell me. So, we can see here that key engagement has been around, you know, personal updates relating to humour, such as farewell, wet wet falls, hello, with go, some funny posts, show high engagement rates. He wrote about some awareness campaigns, save the koala day, and see when I saw you made a comment in the chat before about koalas.
Yes, I think, you know, it's a big thing down here. Awareness campaigns, user interaction, humourous content. And then he also says that he's also posted educational content, all those not as engaging as the humourous posts.
So, it's kind of, you know, it's kind of interesting. And then it says here, would you like, you know, like some detailed strategies to help improve it? And I can say absolutely yes, you know. Actually, I might, do you do the same? I'm actually going to say yes, please.
I don't know, I just feel like when you're talking to these pages, you have to be friendly. I'm not sure it actually really makes a difference, but it feels good. So, I'm going to say yes, please.
And then, because we've been doing analysis of being engaged for the last 10 years, and we've collected so much data, case studies, benchmarking data about what works well. It applies those insights. And it applies what Ben has been writing about.
And it gives me some ideas about what it is that Ben can do to, for this to become even better. And give him ideas about what he'll be posting about. So, I think this is pretty cool.
You know, I'm, as I said, I'm actually, I might be a little bit lazy. I know that I'm going to be sitting with Ben, and we're going to be having a chat towards the end of the year to plan Q1. So, I'm just going to post in here, I'm helping Ben with a content plan.
What can you suggest? And it's kind of interesting, like you can actually now sit here. Because I know what Ben has written about, I can actually come in and say, hey, how about this? How about we sit down and do a plan for the next year? Or at least Q1 next year. You can see on the February, it even says Valentine's Day.
You know, so it's sort of this, it's conscious about what's happening at different time of year. And I guess if you only got five minutes, or 10 minutes, and you're sitting there going like, what do I talk to Ben about? This could be a really smart way of, it's a whole different way of engaging with analytics and making it far, far more actionable. So, these ideas I think is showing the way for how generative AI is really moving from not just the content creation side, but also on the analytics side, and help us understand what is really performing well, what's working well, and identify some of the areas where we can do better.
And then you can use the data and the insights for that other part of, you know, not just the planning and doing, but also the checking and the ethics. I can also go in and then find one of the, like, I might find this top engaging post, and I'm going to click on the little spiral in here. And I'm getting a, I'm actually going to get a, you know, you can imagine some of you guys that are working in large companies, or of course that, oh, now I just closed it just as it was starting up.
Sorry for that. So, if you've got a conversation that might have 200 replies, getting a quick summary of what this means, you know, Ben could be saying, hey, what about that? The launch of some agency I ran, I got 200 replies. What are people saying? And being able to go in and quickly, pretty quickly saying, this is actually what that discussion said.
And to be able to go in and say, like, I know this particular conversation is about Ben moving to Lithgo. We can actually go in and say, what do people say about Lithgo where you're moving to? And you'll come back and then say, well, people have shared, you know, comments about, oh, I've been to Lithgo before, or like, you know, like Blue Mountains or whatever. So, it's pretty cool that you can be able to talk to the data like this.
There's one more thing also, like you can see Ben's post. Some of them have gone well, but like some of them haven't gone particularly well. So, this one only had one reply.
So, I can also click on this one. And for my idea here might actually be to say, what is it that Ben could have done better for this one? Like later on, I'm sort of his improvement coach and I'm sitting here with Ben. So, I'm actually asked a question here to Dr. SWOOP, what could Ben have done to make this more engaging? So, and again, because we've done benchmarking about this for such a long time and connected so much data, we actually have a lot of insights that Dr. SWOOP is fit with.
And you can sort of see like how the generative AI, understanding the data, driven interesting conversations from us as communicators, and can give ideas to people about what they could be doing better. So, I think hopefully you agree that you have a much better likelihood of having a really informed discussion with Ben, or at least I have, when they couldn't talk to him about his content plan for 2025. But I mean, Ben is just one of several people that is on the leadership team that I'm tasked at helping out.
So, I'm not gonna go and see what the other leaders are actually doing. So, I'm just gonna go back and then, now, if I'm like a special anointed person, I have the ability to do benchmarking. And what I'm gonna be doing here, I'm gonna show you a list of people it's kind of a mix of real data and made up data, because we've got myself and Marianne and Laurie, and we are three of the founders in SWOOP.
Ben is one of our senior team leaders. And then I've got Eric in here, who's a relatively new developer. So, they're not exactly leadership team, but I want to show you some interesting data points.
So, again, I'm gonna click on little sparkle icon up here. So, instead of looking at the data, and if you're a bit like me, like, well, actually I do love numbers, but sometimes tables, I don't know, it's just like so much data, you sort of go, what should I really focus on? So, it would be really nice if I can say, well, can you just help me summarize what is the data I'm actually looking at? And it's sort of coming up and saying, well, this is really what we say, what the table is really showing. So, it shows that Marianne and Kai have high levels.
It says here that Marianne, Kai and Laurie have all engaged in personas. And then it says, but Eric, he's classified as an observer, he hasn't got a lot of activity. Now, remember, the deal that I had with this leadership team was that we all agreed that we were gonna be using, we've engaged, it was a part of that sort of handshake we made to improve employee engagement.
So, and I can see that, I guess I can see that Eric hasn't done a lot, and I was gonna say, gee, he's our CTO, what can I do to encourage him more? So, you see, it doesn't stop at just actually looking at the data, it's then going like, oh, then what can we do about it? And they come up and say, well, there's a few things, and even in the presentations before from Australia Post, you heard some of these things, emphasize the importance of what this is doing, the ability to connect with the people, to reach people. And it comes up with a set of ideas to highlight. And even saying, would you like detailed suggestions? Yes, I like details.
I'm not gonna be polite anymore, I'm gonna use the exclamation mark. I can't wait, give me some insights about what I'm gonna do. And there are ideas here for things that we could be doing, or that I could be doing now with Eric.
And one of the things I was mentioning is that I was actually also sitting here and working with Ben on this. By the way, you can see, advice we're coming up with here is based on things that we've learned by doing this benchmarking over the last 10 years that we've engaged. And in the last, I think, last eight of the 10 years, we started to really focus in on what the top performing organizations were doing, and start deliberately to go out and talk to them about what they were doing, how they were engaging senior leaders and so forth.
So the advice that's coming back is basically fed in all this from our global benchmarks. But as I said, I've been working with Ben, and Ben is actually doing pretty well. So I'm sort of gonna say, say, well, what is it that Ben is doing that Eric could try out as well? Because I'm helping the whole leadership team here.
So I'm gonna ask this question and see what Dr. SWOOP says to that. And it's kind of nice when you're sitting with someone, and I think that's where some of that, some of the competition comes in. You sit with one person and say, hey, this is what Ben has been doing.
That actually works really well. Maybe it's something you should try as well. And when you can give those examples about what people are doing, and say, start to compare them, then actually it helps a lot.
And you can go in and set specific goals and help them. So you can see this loop between the data, having a plan for engaging the senior leadership team, pointing out who is doing it and how they're doing it, and then starting on the, so that's the first part of the planning, that's planning the leadership team, we are gonna do this, then actually doing it, helping Ben and Eric with their content plans and their campaigns, checking to see whether that actually worked. And then in areas where it didn't work, and it could be this case that Eric was too busy or just couldn't be bothered, we go down, we go back and revisit the business case with him, we act on the data and then we try again.
So I think that improvement cycle becomes so important at a time we are at the risk of using generative AI to simply pump out more content and not really thinking about how well it's working. Now, does this actually work? Does this thing actually work? Well, I'm gonna show an example of when you get senior leaders involved, and it's a case study that we are very happy to write about and publish in December last year. And it's from Westpac New Zealand, who are also featured in the report.
This is Russell Jones, who is the CIO of Westpac New Zealand. And here, the report is basically, or the blog post is called How to Become an Influencer in a Month. And that's exactly what Russell achieved.
And it doesn't actually take that much. If you read some of the quotes from him, he says here, despite his busy work as a CIO of a major bank, posting once a week is relatively simple. He's had feedback that people are looking for to see emojis and so forth.
And he's really taken to the engage. And of course, the data and helping them understand how it's performing is one of the big drivers of that. So I'm just gonna go back and leave you with a couple of final points before I open if there are any questions along the way.
So let me just go back and just find the PowerPoint slide over here. So I think hopefully what I've shown you today, and I think this is pointing to where the future of generative AI sits for analytics in certain applications. Today, most of us, most of you hopefully, also using, collecting data might be generating dashboards, but you're already now collecting data that helps you understand what people are doing today.
Plan, do, check, act, that's the checking bit. We can do that with dashboards. Now, being able to then understand what those dashboards mean is where is the first, I think, really differentiator.
And that's the middle transition we're seeing here where we're using tools could be Dr. Sewell, can be other things, whatever you have at your disposal to then apply generative AI to interpret the data that you collected. Now, we are in a really fortunate position in SWOOP analytics in that we had benchmarking data for 10 years that allows Dr. Sewell to come up some very smart ideas about what works, what doesn't work based on this carefully curated garden of high quality content that we collected over 10 years. You can almost imagine as I show you the things, the examples with how I was helping Ben improve and how I could help Eric become better, you can almost see that why do I even have to go in to a dashboard and click a button to get this? And I think what this next leap we're gonna see forward with generative AI and analytics will be that the dashboards almost stay in the background.
Actually, I think what we're interested in as communicators is give us the insights. Now, if insights need to be backed up by evidence, back it up by evidence, but we don't have to lead with the dashboard. We don't, might even not have to lead with that table I showed you before.
We might even have to lead with the dashboard that show how things are going up and down in the future. I think there's a real opportunity for us to take the insights and say, oh, would you like to improve how your leadership works? Here are the two things, celebrate what Ben is doing and then work with Eric. Do you want to see some data for that? Here's the data that supports the recommendation.
This is obviously only gonna work if you have really solid data. So for those of you that are sitting there going like, oh, should we start collecting data about how our intranet works and how we've engaged works and like our email open rates? Absolutely. Like there's almost, you have to start doing this straight away because if you don't do that, you don't have a platform upon which generative AI can help you interpret the data.
So, and if we don't do that, like if we don't do that, I think here's the real crutch for us. There's a bunch of different risks that we run. We're flying blind.
We run the risk of generative AI now producing even more content that we actually don't know whether it's working or not. That means that we could keep on making the same mistake, not making evidence-based guidance and recommendations to our senior leaders about what actually works and what doesn't work. And I think we can really only get that COVID seat at the table if we are able to bring evidence and data to the table and say, we recommend you do this for this reason.
And I think that's tied using also, of course, connected to the budgets, the ever-shrinking budgets that I'm sure that, often I hear conferences, people ask, so how many people do you have in the comms team? And there might be two, three, four, five, six people. It's not often they are in the hundreds. And I think one thing everyone is fighting for is budget to get the work done.
Now budget comes when you can prove a business case and you can say, this is why we're doing it. This is the impact and effect it will have. And without data, that becomes really hard.
So I hope that this, I really think there's a huge beaming light at the end of the tunnel here. And I'd like to use the opportunity to get you to think about generative AI, not just for creating content, but also for analysing to make sure that what you produce is actually meeting the needs of the audience.
Meet the speaker:
Cai Kjaer
CEO & Co-Founder
SWOOP Analytics
