Katie [00:04]
Hello, and welcome to the Internal Comms Podcast with me, Katie Macaulay. In our final episode of season 10, we’ll be exploring a topic that has been dominating the headlines all year: artificial intelligence. Now to illustrate one of the many uses of AI, I will be asking our sound engineer Stu to swap my voice for synthetic Katie, at some point in this introduction, see if you can spot when I stop talking, and my generated AI voice takes over.

What does this rapidly advancing technology mean for our profession? How can we integrate AI into our work into our processes? What ethical issues, privacy and confidentiality considerations should we be keeping in mind, and which tools are best for which tasks? To answer all these questions, I went straight to someone who has been at the forefront of technology and communication for decades, the legendary Shel Holtz.

If you have ever attended one of Shel’s presentations at world conference, organised by IABC, the International Association of Business Communicators, you will know that every year it is standing room only. Shel has a remarkable ability to anticipate what’s next for our profession. His knack for seeing what’s around the corner in technological terms means I really couldn’t think of anyone better to guide us through today and tomorrow’s digital landscape. Now, Shel mentions lots of useful resources and AI tools in this episode. In case you miss any of them, they are all listed in our show notes on our website. That’s abcomm.coi.uk/podcasts, you will find a downloadable transcript there too. So, without further ado, it is my very great pleasure to bring you Shel Holtz.

So, Shel, welcome to the Internal Comms Podcast. It is a privilege a pleasure to have you here.

Shel [02:34]
I’m thrilled to be here, Katie.

Katie [02:36]
Normally, I introduce my guests upfront as part of the intro. But this time, I really did fear that no introduction from me would do justice, to your career, to the contribution that you’ve made to our profession. So, for anyone who might not know who you are, I can’t believe there will be listeners that don’t. But can you give listeners a sense of your background, your areas of interest, your expertise?

Shel [03:04]
Sure, I am primarily engaged in internal communications, I should say I am Senior Director of Communications for Webcor, which is a commercial general contractor and builder. We are operating throughout California and headquartered in San Francisco. I’m based in the East Bay of San Francisco. And I’ve also spent a lot of time working on the integration of digital technology into communications. I got interested in that in the mid 80s. And have pursued that fairly relentlessly ever since then. So, for 21 years, I was a consultant working mainly with large companies on both of those things, both internal communications… worked with several large organisations on internal comms audits, and other consulting engagements, but also worked with companies to help them figure out how to use digital tools in their communications. That would be anything from the early days of email up to social media in the late 2010s, which was when I made the move over to Webcor. I was the director of communications at to Fortune 400 companies and spent some time in a couple of global HR consulting firms as well. I’ve written some books, I have my own podcast, and I’ve been very engaged with the International Association of Business Communicators for a long time.

Katie [04:36]
Fantastic Shel. Just a quick supplementary question on that from me, because you have very much worked firmly both sides of the fence as the independent consultant and as the in-house practitioner. Do you have a favourite place to be or do you like them equally for different reasons?

Shel [04:55]
Oh, yeah, I’d have to say I like being back on the client side. There was a lot that I liked being independent, things that I miss. But I like working with the same people to achieve long term objectives, rather than coming in handling the assignment, especially audits, so you’d be in there for six or eight months, doing an internal communications audit, go back with your recommendations about how to improve this, how to move the needle, and then leave. You never really get to see how well it worked not to have a hand in the implementation. Off to the next audit. I do like that long term characteristic of being on the client side.

Katie [05:37]
Yeah, that makes perfect sense. We’re here primarily to talk about the subject of artificial intelligence. But I thought perhaps upfront, we should start with a few definitions just to set the ground for everyone. In layman’s terms, how would you characterise, how would you define, artificial intelligence, and also generative AI? And I was wondering, in terms of that definition, could you give us a bit of history because although we’re hearing a lot about it at the moment I think I’m right in thinking that actually some of this technology has been around for some time,

Shel [06:38]
It has been around for quite some time. People probably remember Deep Blue, the IBM System that won a chess match against a grandmaster. But that’s machine learning, which is a subset of artificial intelligence. AI is basically when machines act like people… do things that that we see as requiring human intelligence. Think about identifying something that you see or making a decision or translating something from one language to another or analysing data. And what AI does, essentially, is predict. That’s all it does. It does not reason. It has no reasoning capabilities. It is not actually intelligent, it just predicts based on your query, what’s the next thing that it should deliver to you. In terms of generative AI, it’s what’s the next word. Hence, is why you’re never going to see wholesale pieces of some content lifted from another source and delivered as a response from a generative AI tool… is that it’s not sampling from a dataset, it is actually predicting the next word to give you – word by word to create a new sentence, a new paragraph, a new image. We’ve been using AI for years. If you watch Netflix and see the recommended shows for you, AI produced that. Same thing with Amazon when it gives you shopping recommendations or reading recommendations. There have been organisations that have been using it for years. I think it had to be seven or eight years ago, I read about a Japanese insurance company that laid off its claims adjusters, because it found that the AI was able to look at the images of the damage to the car and assess what the claim should be. They didn’t need a person to do that. And that’s been a while.

What happened was a year ago, in November of 2022, Open AI released Chat GPT 3.5, a generative AI tool that was open to the public. So, you could just go in and enter text into the text prompt field and get a result. And that was, for a lot of people their introduction to AI, it was the first time they’d ever used it. I’d been using a tool called Jasper for a few years before that, and it’s a generative AI tool. But this one was free, so everybody could pop in and use it. And then Microsoft invested $10 billion and made it part of Bing. Bing had been a runner up search engine that was the subject of some derision before this. Suddenly, it became a viable option as a Chat GPT and eventually Dali the art generator was woven into it. So that was the first-time people were exposed to it and started really talking about this idea of AI.

Generative AI is simply AI that produces a response from a prompt… it generates a response. They can produce text from a text prompt, music from a text prompt, images, video. You can produce text from an image. These work in many different directions. In addition to Chat GPT there’s Claude, Perplexity Google’s Bard, Midjourney, Stable Diffusion – all examples of generative AI.

Katie [09:51]
I want to come back in a moment to talk about some of the tools that you find yourself using most, the ones that you particularly find helpful. But when we first talked about this and I think it might have been on actually the FIR podcast, your podcast, I heard you mentioned AI as being a force multiplier. What do you mean by a force multiplier?

Shel [10:13]
Force multiplier is generally a military term, it’s used to describe the effectiveness of power that comes from adding something to something. Advanced weaponry is an example of a military force multiplier. So, generative AI does the same thing for workers. It’s a capability that gets added to an individual’s or a team’s toolkit, it gets woven into their workflows. And that results in an increase or an improvement to output or effectiveness or using a tool like Chat GPT can improve your productivity or your performance. In fact, I saw – and this was in the early days after Chat GPT was released – I believe this was a professor who analysed the productivity improvements of workers using Chat GPT, just individually, not using any kind of mandate from the company or policy or specific procedure, they were just using it as an assistant. And their productivity increased. If I remember this number correctly, it was 35%, which was more than worker productivity increased with the introduction of the steam engine in the industrial revolution. So that’s considerable. And that’s why I see it as a force multiplier.

Katie [11:34]
Is that how you’re using it mainly in your day-to-day work? So, coming on to talk about how you’re using it and the kinds of tools and platforms that you find yourself using…

Shel [11:44]
Mostly these days – yes. And that’s just because I really haven’t had the time in my job to analyse all of our processes and workflows and start to map out where to add AI to those workflows. I think that’s where we’re going to see the real value from this. But yeah, I use it every single day. Seriously, not a day goes by that I don’t use at least a couple of the tools available. I’m paying for a few. I pay for Chat GPT, the pro version. I pay for Midjourney and Stable Diffusion. Lately, I’ve been really enamoured of Perplexity, I think this is going to pose probably the most serious challenge Google has seen to its dominance of search. I really have not needed to use Google since I started using Perplexity. Rather than give you the 10 search engine results on a page that you have to scroll through to see if what you’re looking for is there and then maybe click through a few pages and go ‘Nah, that’s not quite what I was looking for’, and then go back to Google. You get a narrative answer in Perplexity that is based on a search. And it includes the links to all of the resources that it used to produce that result. And it’s spectacular. You want to talk about a time savings when you’re just looking for information. So, what is this going to do to search engine optimisation? What is this going to do to the whole approach that we take to finding content on the web?

And this leads me to a few worries, because a lot of organisations that produce content, rely on Google to drive traffic to their sites so that they can monetise that effort. And if what you’re getting now is just the answer you need without a reason to need to go visit the site, what’s the incentive for people to create that content in the first place? So, I think we’re gonna see some interesting evolutions of what the web is and what digital content is over the next few years as we see this play out.

I’m using Midjourney pretty heavily. Not as much as I was before, not since Open AI released DALL E3. Which is amazing and far easier to use. And Stable Diffusion just unveiled a suite of image tools compete with Adobe Express. There’s a tool for removing backgrounds, for changing backgrounds, for changing the sky in an image. There’s an image upscaler. All of this in addition to the ability to enter a text prompt and get an image in return.

Katie [14:27]
Just so that people know – Midjourney, DALL E3 as you say, these are all image generators. I’m certainly using Chat for… certainly every day, I’m not sure how many times a day. I’m finding at the moment though, I would say it’s useful as a jumping off point rather than have the blank page in front of you. You put the right prompts into chat, and it will just give you that head start it will give you something to work off. It’s almost like giving something to an intern and get something basic back but it moves you forward a little bit. So, I’ve heard this phrase, don’t worry about AI taking your job, worry about the person who knows how to use AI, taking your job. Which would you feel that’s a fair assessment of where we are today?

Shel [15:13]
Without any question. AI is definitely going to cost some jobs. New technologies have done this forever. There’s a label for this called ‘creative destruction’. There’s not a lot of buggy whip manufacturers out there anymore. But it’s going to end up creating a lot of jobs. And it’s going to end up being a common tool used by knowledge workers in every industry. When I heard that schools were banning it initially, because they were worried about students cheating, I was very worried because what you really need is schools teaching students how to use it, because the expectation in the workplace, by the time they graduate and enter the workforce, is going to be that you know how to use these tools. Also, these tools are going to be woven into the other tools that we use. I think we’re going to reach a point where you don’t go to an AI, just the tool that you’re using – whether it’s Microsoft Word, or Adobe Photoshop – they have the AI built into it, and everything will.

Katie [16:19]
I’m noticing all the things that I pay for Grammarly, for example, all of a sudden, there’s this button that you can press that will re-word, the whole paragraph for you. And that’s just popped up. So, I’m imagining exactly what you’re saying, it’s going to be woven into everything that we use every day. Everything we use every day, we just get smarter, I’m guessing.

Shel [16:38]
I don’t know, if you’ve seen the website, There’s an AI for That. But it lists tens of thousands of AI tools that have been built mainly using Chat GPT as its foundation. But they’re categorised, and they have a newsletter that talks about the new tools that have been uploaded. Yeah, There’s an AI for That is a great site to keep an eye on.

Katie [17:03]
Great handy hints there. Thank you very much. My observation is that if we think about this through the lens of internal communication, I think, as a discipline as a profession, we have slightly lagged behind in adopting new tools and new technologies. I think we were frustrated by our channels for many years, for example. Finally, they’ve become two way and social, and all the rest of it, but it seemed to take some time. I don’t know if you would agree with that analysis. And if so, why you think it’s been like that for so long.

Shel [17:39]
I think that’s historically true. And I think it goes beyond internal communications. I think this is true in public relations, as well. And I do work in PR right now and I have in the past. And, I’ve complained about this, I’ve written about it. Our counterparts in marketing and advertising tend not to lag the way that the communication side of the profession does. And I think there’s a couple of reasons for that.

One is that their budgets are higher, and they have the wherewithal to invest in new technologies. They are more inclined to experiment, because they’re trying to find new and innovative ways to attract consumers or b2b buyers. And partly, I think it’s because they don’t feel any pain. In the PR and communication side, we tend not to change until we feel pain. If you consider, for example, the web. When the web was relatively new, PR agencies were saying why do we need to get involved with the web? Our clients aren’t asking us for it. It’s not replacing other things that we do and our billings are high. So, why should we do this? And it wasn’t until clients started hiring boutique agencies that specialised in digital communication that the legacy agencies started to say, well, we’re losing business. And we really need to learn this.

I remember, I was hired when I was an independent consultant by one of the world’s biggest PR agencies to come in and help them get their minds wrapped around how to do digital communication. Because now their clients were expecting it. I think those are the reasons that that we have lagged with these new technologies. With AI, these really aren’t considerations. These are low cost or no cost. But one thing to keep in mind is that most of these tools are on an individual basis. So, it’s not like you’re acquiring an enterprise tool. Chat GPT is 20 bucks a month. And I haven’t asked my company to pay for that. I use it for non-Webcor stuff. So yeah, I pay for that. I think it’s more a combination right now of fear and a lack of understanding about how to use it. Mostly, it’s interesting, I hear people talking about writing. With AI, ‘It can write this, it can write that it can adopt this tone, it’s not quite as good as a real human, oh, it’s good enough for the stuff that we do, that just needs to be good enough.’ But for me, the fact that it writes it, and I’ll confess it, I have used it for that a bit. But that’s the least interesting thing about it from my perspective. Whereas some technologies require department wide adoption, let’s say the use of desktop publishing for print publications, if you’re moving to desktop publishing, the entire department is going to be doing that. Or the use of HTML to produce online content… But AI is more a personal assistant. You use it one on one, and by internal communications, its use is going to increase when there’s greater understanding of the benefits of using it. And I think we’re also going to need to see this unreasonable fear evaporate eventually.

Katie [21:12]
But when you say basic writing, that’s the kind of least interesting… the more interesting one being then productivity, speeding things up, a massive amount of analysis of a big amount of data very quickly. I have to be honest, just as a, as a trial, someone asked me to review their book, hundreds of pages, as a PDF, attached it to chat, and said, produce a book review of this, I had not read the book. Now, of course, I am going to read the book, I really will. But I was just curious to see what it would give me and again, that you can look at that and say we’re 25% ok. I wouldn’t be happy enough to put that out in the world with my name on it. But it gives me that sort of starting point. That is how I’m using it at the moment. So, what in your view, is that the sensible thing to be testing and trialling and adopting at the moment?

Shel [22:05]
Absolutely. With Claude from Anthropic, you can take multiple documents of different types and upload them as part of a single session with the AI and start asking questions about it and getting summaries and getting kick started on a synopsis of what all this means. So, if you’re working on an article, and somebody says, ‘Here’s some background material,’ and you don’t want to read all of this. Have the AI do it for you and tell you what the salient information is that’s in there.

But I think there’s more I think you can use it to develop communication strategies. I’m not saying you want to replace your experience and understanding of the organisation and the conditions. But it can certainly produce some ideas that you hadn’t thought of. One that I’ll do is I’ll develop a strategy for a particular initiative, and I’ll give it to Chat GPT. And I’ll say, this is the strategy that I’ve developed, you will see in the strategy that my objectives and goals are A, B, and C – what am I missing? And what is misguided in this strategy? And it’ll tell me. I don’t necessarily use all of the feedback that it gives me. But almost every time I get something out of that I hadn’t considered. So that’s incredibly valuable.

I also use it to do things like subheads, I can give it a long article and say, ‘Please write descriptive subheads about every three paragraphs’. I can use it to produce a four-bullet summary at the top of a lengthy article, give me headline variations for this article. There’s just all kinds of things that you can use it for. But analysis, for me, is really the big thing. It’s a wonderful research tool, as long as you verify the facts that that it gives you because they are prone to hallucination. So, you do have to Yeah, you do have to fact check. That’s part of the workflow. Right?

Katie [24:15]
Exactly. Don’t forget, you have to verify your sources and your data. I suppose we have to deal with the – I don’t know if it’s the elephant in the room question… But there is a spectrum of opinion about AI. One side of the coin you’ve got this is a digital superpower, that ultimately, we’re not going to be able to control, it’s going to end up controlling us when it is fully developed, fully sentient if it ever can get to that stage. Versus the view on the other side, which is, look, we’ve seen many technological revolutions over the years. This is just another one. I’m just wondering where you stand.

Shel [24:51]
The worry is that you’re going to get to this singularity artificial general intelligence, which is the term the tech scientists are using to describe artificial intelligence that can think like a person. Right now, it can’t, right? It’s not even close. There are people who think it’s not even possible to get there. Others think we’re 50 years away, some people think we’re already there. So, it depends on who you talk to.

I am not worried. I think Hollywood has embedded this vision of the future of AI in our heads, and it’s Skynet and the apocalypse and The Matrix. And I can see when we do get to AGI, and it can perform multi-step tasks, where you say, ‘Look, this is the task to achieve, go do whatever steps it takes to achieve that task.’ That if it’s interconnected to all kinds of networks, that it could be dangerous. But I think that there are a lot of people working to ensure that never happens. In fact, the whole brouhaha at Open AI where Sam Altman, the CEO, was fired and then rehired, was all based on arguments about rushing forward with right advancements in AI versus being very safe with it. But imagine that somebody gave an AI the instruction to end climate change. And as one of the steps that determines that people are responsible for climate change, so I gotta get rid of all the people. That’s the type of thing they’re working to make sure can’t happen. But we’re nowhere, anywhere near that type of capability with AI right now. Right now, it’s a productivity tool. Its nefarious uses exist, but they are around bias and disinformation, and things like that.

Katie [27:55]
We hosted a webinar not that long ago about the explosion in all audio AI tools. And they are quite amazing. So, we demonstrated how you can have your chief executive, for example, give his end of year message in his own voice. But speaking very much language he cannot speak. That might be Mandarin, that might be Spanish, whatever it is. Someone in the audience immediately picked up on the fact that this is slightly dodgy in terms of not being that ethical. There were issues around authenticity for her. And I just wondered what you’re thinking is on this. Is it all about disclaimers and caveats before we rush off and do all these amazing things that the tools that enable us to do?

Shel [28:40]
Yes, it is all about disclosure. This is synthetic media – synthetic media is not new. Deep fakes were a thing well before Chat GPT was launched last November. And they’re getting easier. I submitted my speaker proposal to IABC for the conference in June in Chicago to be about synthetic media, both how we can use it and what the challenges and issues are. And when you submit a proposal to speak at IABC you have to attach a 60-second video pitching your proposed session. And I did and at the end of it, you hear me say ‘And by the way, this video isn’t me. This is synthetic media. It is an avatar that I created. And it’s just reading the script that I was admitted to it.’ You can’t tell that it isn’t me. This is how good this stuff is getting.

Now there is value in having your CEO deliver a talk to employees – whether it’s audio or video – in their native language and be able to look at them in the eye as it were over a video screen. And I think it’s fine. I don’t have any issue with that at all. As long as you start that by saying, ‘Look, our CEO does not speak Portuguese. But we have this AI tool that allows us to take the speech that he delivered in English in our corporate headquarters in Connecticut, and have it presented to you in your language. And because we can do that, we did it as long as people know, you’re fine.’ I don’t see any reason that we shouldn’t take advantage of tools like that.

Katie [30:26]
I totally agree with you. As you say, it’s just that disclaimer, that disclosure upfront, I think. I know, some clients, we’ve tried to push the boundaries on some stuff that we’ve delivered to clients, there’s been AI generated. Say, for example, we’re talking about a new initiative inside an organisation that happens to be an AI initiative, we’ll have artwork to support that campaign that’s AI generated, because it seems like a cool thing to do to showcase the technology with artwork that’s actually produced by the technology. But then the client’s pulled back from that and said, ‘Oh, no, hang on a minute. We’re not quite sure how and why we’re going to be using AI with rolling back from it. There’s concerns about confidentiality and data privacy and all the rest of it.’ Can you talk us through what we need to be aware of as IC professionals in terms of what might be the governance that we’ve got to consider as we’re using these tools on an individual basis?

Shel [31:23]
I think those are two legitimate issues, privacy and data protection. But I can’t remember another… and that’s not true, I can remember a technology that led organisations to forget that they knew how to conduct a risk benefit analysis. And that was blogging. Because what would happen was because blogs were done quickly, and somebody would write something, and then get in trouble for something that they posted. So maybe it was a securities violation, and the company got fined as a result of that.

Now, they’ve written 50, blog posts, and those blog posts, if they were done strategically, reduce the number of calls to the call centre, led to improved sales lead to greater understanding led to support for an initiative… all of these can be measured. And you have the one that ended up costing the company a fine from here in the US the Securities and Exchange Commission. So, the lawyers say, ‘We can’t do that shut it down.’ Well, if you look at the value you got from those other 49 posts, and it was 10 million, and your fine was 2 million, you were $8 million dollars to the good, right? And yet somebody said shut it down. That’s dumb. You’d need to do the risk benefit analysis. And if the benefit outweighs the risk, then it becomes your job to ameliorate the risk. And you do that by basically training people. We trained them on other tools where there were risks – think about dangerous machinery, right, we train workers how to work around dangerous machinery without getting hurt.

So, do the training and the related communication so employees know how to protect privacy when they’re using these tools. Establish policies about the details that you shouldn’t share in a prompt or in something that you’re uploading. But train people also on the settings on these tools, because both Bard and Chat GPT have settings that prevent prompts or answers from being maintained on the system. So, you can go ahead and share anything you want, as long as you’ve… you know, set the setting that way. But I would not deny the productivity here that the company can realise, especially when you stop and think your competitors are figuring this out. Competitors are figuring out how to be more competitive using AI, while you’re busy denying its use because you’re worried about some of the possible misuse and the consequences of that.

I’ll tell you… Open AI has these GPT builders available, where you can build your own. So, I took our employee handbooks, PDFs, we have two we have one for our salaried employees and one for our hourly. And I uploaded them, and I gave it the instruction that it should be able to answer any question the answer to which is contained in these documents. You should be able to delineate between an answer for an hourly and an answer for a salaried employee. You should be friendly and casual but never unprofessional in your answers. You have to have a paid subscription to use these things right now so it’s not like I can make this available in the company. But right now, if somebody has a question related to their employment, they have to first go find the PDF of the manual and then page through it to try to find where the answer is. Now you just go to this text field, and you go, ‘I’m not feeling well, who do I need to notify that I’m taking a sick day?’ And it answers. And it’ll say, by the way, this is true for both salaried and hourly employees. I asked at once, what happens if I show up drunk on the job? And they gave a great answer. It’s crafting its own answer based on nothing except what’s in the text of these documents. It’s not confabulating, it’s not going out searching, or using anything from its dataset. It’s just what you have provided it.

One of my favourite ones that people created was you list ingredients that you have in your pantry in your refrigerator, and it will give you recipes that require you to use those to help you get rid of stuff that’s getting old. You can imagine the productivity enhancement from not having to search for the employee handbook and then search for the answer. And the employee satisfaction and improved employee experience that comes from just getting the answer to the question you want. Now project that out to all of the information and knowledge that’s contained in all of the organisation’s networks. And just let people ask the question gets the answer. What’s our best practice for this? Well, how do I go about doing that? And just the answer comes back. And it becomes a single source of truth to you, you don’t end up with two documents that have somewhat different answers.

Katie [36:31]
That’s wisdom of the crowd, really working. It’s funny, when you talk about giving employees permission to play with these to try these out. I can actually remember… we must be going back 15 years, or maybe a little bit more. I can remember having conversations about whether knowledge workers, so people sat in front of a screen all day, should have access to the web. Like if we give them access to the web, they’ll just be looking at cat videos all day. Can you imagine now that conversation happening? It just wouldn’t. So, I’m with you on this completely.

Shel [37:08]
I used to have a blog called Stop Blocking. And the stories about organisations that were blocking access to the web and the consequences of that. I was actually at a client, I won’t name them… But I was out at an engagement with this client the day that it blocked web access. People had it and the decision had been made that we shouldn’t allow them to have this, and they blocked it and the phones started ringing. ‘I use this website to get this data for this weekly report. How come I can’t get to it?’ Just dozens and dozens of calls like that. So yeah, we do tend to see organisations jerk their knees in response to these things rather than take a reasoned analytical approach to them.

Katie [37:55]
I think on FIR podcast, you raise this idea that it might not be that long, for some of us… at least might be reporting to an AI boss. And I thought, ‘Yes, I can actually imagine that. That’s not that weird, actually, when you think about it.’ If I fast forwarded us, I think 10 years is almost impossible to say. But maybe five years. Do you have any predictions, particularly from a commerce perspective, what we might be seeing what the landscape might look like?

Shel [38:30]
I think the landscape will look a lot like it does now. I think we will just be using these tools more, which means that we will be spending more time on the high value dimensions of our work and less time on the drudgery that the AI is able to take on for us. Now will we be reporting to an AI boss? Maybe there’s actually an AI CEO out there. Now her name is Mika. She is a research project from a company called Hanson robots and a rum company in Poland called Dictador. And they’ve customised the AI CEO to represent the company and the company’s unique values. Mika has a wide range of tasks like helping spot potential clients and selecting artists to design bottles. She uses a decision-making process that relies on the language model, right, the extensive data that she has been given access to and trained on data analysis and aligning with the company’s strategic objectives. She has no personal biases, which means that the choices that she makes are purely strategic. It’s worth noting that major decisions are still made by human executives though. Does Mika have all of the responsibilities of a traditional CEO? No, it’s more of an experiment. But if you think about the CEOs role of looking at the big picture, and taking in all of the information and guiding the organisation in the right direction, it seems to me they’ve got the right idea with how they’re taking advantage of this, this technology in this way.

Katie [40:16]
You can imagine can’t you, often think as a chief operating officer as a CEO, maybe even as the finance director of these big global organisations, so much of it is about surveying the landscape, all the inputs of all the data, all the insights that you’re seeing… And being able to create that big kind of view of that and see what’s important, what’s not important? Where do we go next forecasting, all of that sounds…

Shel [40:40]
I think about the industry I work in, and you’re trying to figure out what kind of projects do you want to pursue, given the economy and the capital plans that you see coming out of some public organisations and trend data and all these different factors that that are going to have an impact on what kind of buildings are going to be built over the next five years. And to be able to look at a large language model that has all of this data and then give it more that you have to help you identify the direction to take… And what types of markets to pursue. Should we be public? Should we be private? If we’re going to focus on public work should we be looking at infrastructure? Should we be looking at hospitals? Should we be looking at student housing or life sciences? Which baskets should we be putting our eggs in? That kind of analysis… I think AI is going to really up our game on our ability to make good correct decisions.

Katie [41:51]
Do you think there thought that if we are thinking about, you know, people entering the job market, and the values and the skills that we’re going to need to develop successful careers in the future – are we going to need two things? So, we’re going to need the ability to understand how to get the best out of these tools. But we’re also, I think, going to need to become very good at the human stuff. Sort of intuition, imagination, creativity, all of the human-y characteristics. That’s not a word. But it feels like we’re going to need you to be able to do both of those things. Would that be fair?

Shel [42:32]
I think that is fair. I think, again, if you don’t know how to use these tools, you’re going to be at a severe disadvantage on the job market. And not just in communications, pretty much any white-collar job, probably some blue-collar jobs as well. But you’re right. If you look at a profession, like nursing – nursing is going to be dramatically affected by AI. AI is going to identify the correct course of treatment. And in fact, we’ve seen multiple reports of AI correctly diagnosing condition that doctors were unable to diagnose.

But an AI cannot do the human part of nursing, the comfort to the handholding, the encouragement, the compassion, and I think when you look at what we do as communicators, it’s the exercise of judgement. And AI has no judgement at all. It’s the application of empathy around certain things that we need to communicate internally. AI has no empathy. AI has no life experience. So, you have to be able to bring those human characteristics… And I’m thinking, especially life experience, and empathy and compassion, those types of things – and judgement – is just going to be huge. And I think that’s where you’re gonna see, the partnership between people in AI is each one tackling the elements that they’re best at in order to achieve a common result.

Katie [44:08]
I think that’s quite reassuring. The only other thought I had, and this has frustrated me for years that you can work for an organisation for 30 years, and still get a ‘Dear colleague’ letter, but not even addressed to you. And I’m just wondering whether AI might get us to a stage if we can get it working in the HR space in the sort of talent and people space where we might be able to start communicating with individuals we might be able to say ‘Hi Shel, You’ve been working for us now for 365 days, just wanted to say…’ Something very targeted, something very appropriate, something very relevant to your interests, your skills, etc, etc. Can you see that possibility or am I just really flying a kite there?

Shel [44:56]
I think that possibility exists now. If you read the analysis of AI capabilities in HR and communications, and especially in marketing, what you see is there’s a lot of personalisation. I got a demo from Aaron of his tool called PRofit, that’s PR, capital P capital R PRofit. It’s a media relations tool. Now, let’s say I want to pitch the media on a story, and I want to do individualised pitches to reporters who are likely to be interested in covering this particular story. But what I would do is I would start with a database like Muck Rack and find reporters who have written about this. Then I would take hours, maybe days to read articles that they have written so that I can craft pitch letters to them that are going to appeal to their specific interests. What PRofit does is all that work for me, basically, I say find reporters who have written about this, read what they have written, and then craft pitch letters for this particular story that will appeal to them. And then I can tell it to make these pitch letters funny or serious or professional or chatty. And of course, I get to go and edit them before they get sent. But look at the amount of time that saves. And the end result is each of the reporters that has been identified as being someone who is inclined to write about the subject that I have to share. Each of them is getting a personalised pitch letter based on their interests and the stuff they’ve written in the past this capability exists now. Just need someone to develop the tools for HR. HR is already using AI to do things like scan resumes to find the best fit for a job. There are a lot of HR tools that use AI now,

Katie [46:55]
You can imagine that sort of profit tool being used internally to find the champions about a certain subject, because these are people already talking inside the organisation – posting, commenting, liking stuff about that particular subject. It makes it so much easier. Yeah, very exciting. Before we head over to those quick-fire questions, I’m just wondering, how do we get to be more like Shel? So how do we stay up to date? How do we stay current? How do you do that?

Shel [47:28]
Yeah, I read a lot. I subscribe to a number of newsletters. And I’m probably adding at least one a week to the list because somebody will say, ‘Oh, you’re not reading this one, you should be reading this one!’ And a lot of these newsletters curate content from around the web. So, they will link to articles and I will be able to scroll through those and find the ones that are of interest to me or aligned with the work that I’m doing. I also listen to a lot of podcasts. God knows there are tonnes of AI podcasts that have appeared over the last year. Some of them are pretty good. There are also general tech podcasts. One that I’ve been enjoying lately is called Hard Fork. With Kevin Roose, from the New York Times and Casey Newton from Platformer. It’s very entertaining and funny, in addition to being enlightening. But that’s really the trick is… I just read a lot. And when I find something that I want to be able to refer to again, or maybe I see it as fodder for the FIR podcast. I save it to my link blog. Here it is almost 2024 and I have a link blog. It works for me. It’s a Tumblr blog. So, I have the bookmark when I see an article, and it could be one that came from one of the email newsletters that I get. I’ll just click the bookmarklet and save it to my link blog, which is open to the public. If anybody’s interested in what I find interesting, you just go to my link blog, and you’ll see the articles that I’m bookmarking.

Katie [47:34]
We’ll make sure all of these links to everything you’ve mentioned are in the show notes. So, your link blog and obviously the FIR podcast, which has been going, Shel, I think… how long?

Shel [49:15]
January 5, 2005. We actually distributed a test audio file, I think it was December 29, 2004, just to make sure we had the tech working before we posted an episode. There were about 400 podcasts when we started and none on communications. We were listening to stuff like the Daily Source Code from Adam Curry, the Dawn and Drew Show, Geek News Central which is still going yeah. We’ve been doing this a long time. The one of the oldest, certainly the oldest, communications focused podcast.

Katie [49:48]
Fantastic. We’ll make sure all those links are there.

Katie [50:42]
What trait or characteristic do you possess that, above all others, do you think has most led to your career success?

Shel [51:41]
I would have to say curiosity, which is certainly what’s driven my interest in technology.

Katie [51:49]
Just on that curiosity… I can’t quite remember who it was that said this the other day, but that you can make curiosity a habit rather than thinking of it as just you happen to have been born a curious person. You can actually make it a habit. And I’m curious about whether things like the FIR Podcast, encourages it, curiosity, because you have to put some content out in the world.

Shel [52:13]
Absolutely. And this is why I maintain the link blog primarily is I need stuff to report on FIR. This could be one, this could be one. I end up using that link blog for other things. But Neville and I occasionally say, ‘Are we going to keep doing this?’ And the answer is yes. And it’s for two reasons. One is we enjoy each other’s company. And if we didn’t do it, we’d probably talk to each other once every couple of years when either he and Laura got to the US or Michelle, and I get to the UK. And the other is, that it keeps me on my toes looking for this content that will appeal to our audiences and what people expect from us is what’s new. What’s around the corner?

Katie [52:56]
Yes, yes, is exactly the same as my Friday Update. It’s only five bullet points long every Friday, but it forces me to go out there that week and think new reports, new campaigns, new studies, the next big event. So yeah, it’s very similar, I think. So how would you complete this statement? World class, internal communication is…

Shel [53:20]
…More than content. I think there are a lot of people who see communication as a content driven occupation. And there is so much more that we could be doing. I don’t know how many internal communicators have even heard of, no less conducted, a process audit. And we’re not talking about communication processes. I’m talking about the key processes that are used by employees company wide. Because processes send messages and in many cases, the messages they send contradict what the leaders are standing up and telling employees at the town hall meeting. And that’s a disconnect that causes lower engagement scores, lower productivity, higher turnover.

And you don’t know that these gaps exist between say and do until you go look for them. It’s not what somebody is saying that’s contrary to, say, one of the company values, it’s just baked into the way we’ve been doing things for 30 years. Listening is another key communication activity that is not content driven. But I think we need to… It’s not to say that content isn’t important. But it’s not the be all and end all of communication. Communication is happening in an organisation whether we’re there or not. And I think if we are really good at our jobs, we’re helping the organisation communicate more effectively, and that’s beyond articles on the intranet.

Jim Shaffer likes to tell the story about a client he had… a logistics company that was experiencing a lot of breakage, from packages that were being moved out of a distribution centre. And they communicated every way they could think of how important quality was… they had to stop delivering packages to customers that had broken goods in them. And nothing helped, their breakage rate just continued to be high. And they brought Jim in. And he went and did a process audit, he came back, and he said, ‘You know, that people are being paid bonuses based on productivity. They’re making money based on how many packages they ship in a day. As long as you pay them for speed, they’re going to going to ignore your messages to slow down.’ And they went, ‘Oh, I hadn’t thought of that!’

Katie [55:46]
I love that answer. Thank you. What’s the one piece of bad advice or unhelpful thinking that you think consistently as dogged to the communication profession.

Shel [55:59]
That our goal is to earn a seat at the table.

Katie [56:06]
That’s such a great way to end this series. Where do we start with that? There are so many people that want a seat at the table. I don’t know how big this table is. That’s one of the problems that I’ve had with that phrase.

Shel [56:23]
From a communications perspective, there’s only one, if there’s one, it’s going to be whoever’s in the role of the chief communication officer, whatever equivalent that is. Does that mean everybody else who’s communicating in that organisation is an abject failure? That’s ridiculous. But for me, I want to get the call before everybody convenes at the table. I want to get called into the office saying we’re going to the table, and I need an answer to this question, or I need a proposal for how to deal with this. I want to be the one they call not the one who’s expected to show up at the meeting.

Katie [56:59]
Yeah, I love that answer. Thank you very much. And finally, before we let you go, last time you appeared on this show, and I think it was in the deepest, darkest days of COVID. We asked about your billboard message. So, what you would put on a billboard for millions to see. And you gave us a Grateful Dead lyric. You said, ‘Once in a while, you can get shown the light in the strangest of places. If you look at it right’ A great call, I think to comms pros to look deeper and further into their organisation for that great story. Three years on, do you have a new message for your billboard?

Shel [57:38]
Yeah, and it’s not a Grateful Dead lyric, although I was considering finding a relevant one. Always be moving the needle. Always be adding value. The work you should be doing should be meaningful to the organisation. What have you changed in the organisation? The one thing that makes me feel the proudest, after six years at Webcor is people talking to me about how much better things are because of the communication efforts. Since I got here, it was easy, there really wasn’t an internal communication function to speak up before I got there. So, the bar was pretty low. But it still makes me feel great that people think that it’s a better place to work because of the efforts that have been made through the communications function. So, you know what needle meat needs to be moved. This means that you need to understand the business goals and challenges that the leadership, people losing sleep over. And then, you know, what strategy means, what being strategic means is the outcome that you want to produce with your communication is going to help them sleep better, you’re going to be able to measure that, that you’re going to be influencing those things that matter to those people not just reporting on whatever. I have a colleague who thinks internal communications is just corporate journalism reporting on what’s going on in the organisation. I think that’s wrong on a number of levels. First of all, it’s all reactive. There’s nothing proactive about that at all. But second, we could be reporting on things that may be interesting, but don’t matter.

Katie [59:19]
Yes, exactly. Right. And we’ve moved forward so much when I started in this profession, 1990 it was full of ex journalists and then there’s nothing wrong with that. And I’m an ex-journalist. So, I’m not complaining. We know we’re particularly good around storytelling, a nose for story, how to interview people, all of that stuff. But as you say, so much of it is not just about creating great content. It’s the right kind of content in the right places at the right time aligned to the right objective. So, I couldn’t agree more. Shel, thank you so much for appearing again on our show.

Shel [59:56]
My pleasure and it was fun.

Katie [1:00:01]
So that is a wrap for this episode of the Internal Comms Podcast. And for season 10 of the show. I just want to take a moment to say a huge heartfelt thank you to everyone around the world who tunes into this show. Your support means the world to us and this show would be nothing without you.

And using the wonders of AI. Let me reiterate my thanks in three languages I do not speak fluently at all – full disclosure: German Spanish and French

Season 11 kicks off in January 2024. And you could help us choose the guests and topics we cover. We are organising a special LinkedIn Live event in January, where I’ll be answering your questions about the show and asking for your comments and suggestions for future episodes. You can find all the details in my Friday Update and on LinkedIn in the coming weeks. So lovely listeners until we meet again, have a fabulous holiday season. Stay safe and well. And remember, it’s what’s inside that counts.