We recently got the chance to chat to Adam Hirsch, former COO at Mashable and current senior vice president of emerging media and technology at Edelman Digital, about AI in PR.
We took this opportunity to grill Adam — a bonafide tech expert — about the whole spectrum of AI and how it’s affecting the PR world right now. We chat about how popular AI is in PR at the moment, the difference between automation and AI, and how to find out what tools can help PRs and their clients.
Have you found that AI is a hot topic with your clients at Edelman?
Definitely, yes.
AI overall is not only a hot topic for us, but we have an AI centre of excellence that has a number of members around the world. There’s a lot of noise right now surrounding AI, and there are tons of large clients that use AI technologies or are powered by AI, as well as clients whose whole presence is about selling AI.
Let’s go back to basics… what is AI?
The way I like to explain it is it’s basically machine learning, so you take a set of inputs which would normally take a computer a couple of minutes to process and then you’d be able to analyse the output, put it back in, and go again.
Machine learning takes that mindset and starts making decisions for you; it basically learns for you. Of course, humans can intervene and put it back on the right path if it veers off, but machine learning is ultimately what’s at the core of AI.
Once you move past that, you get to decision intelligence.
How does the machine then start to make decisions to analyse that data and figure out what needs to be done next and then go and actually do it?
At the same time too, it’s the ability to make decisions and go above and beyond, and that’s where the biggest opportunity is.
Is machine learning going to take over the PR workflow?
Has it? Definitely not.
Will it long term? Maybe.
The reality is storytelling is definitely not a craft that AI can do just yet.
Can I take a set of scores about a game last night and automate a basic report? Of course. The ability to do that has been around now for 5–10 years. That kind of basic AI — which is used to understand the context of basic data points and then stitch them into a story — is where we’re kind of seeing it the most at the moment.
Things that can be repeated could soon become more and more automated. But I think a lot of people who read a newspaper still want to hear the personality of a specific person, so I still think we’re about 50–10 years from seeing AI take over storytelling.
There are probably many points in the PR workflow that AI is having an impact on, and I imagine there’s a lot of points already where people may be using AI but not really realising it. Would you say that’s true?
Exactly. I actually had the task of understanding all the technologies we use and finding the points where AI exists in those tools and platforms. I then broke them down internally and for our clients so I could tell them what AI truly is and how we’re already using it.
I break this down into three areas:
- Audience
- Content
- The intersection between (like audience behaviours, for example)
The audience part is all about understanding psychometric behaviours and other data points that are harder for people to analyse at a high volume. Content is where we’re seeing some good traction when it comes to analysing content. There’s obviously still a lot of manual work that’s being done in this area because, for example, every client has a different way they want to categorise content.
We’re starting to see the beginnings of automation in understanding content with things like video recognition and text recognition, so we’re there from a basic standpoint.
Do you mean this in terms of when we’re going out and searching for this content? Or when we’re productively pushing our content?
At the end of it, actually; in the analysis part. Search engines have been good for finding content for a while now. Understanding the analysis, like finding out how the audience interacted with a certain piece of content is where AI can have the most impact. Ideally, you want to try and find patterns in behaviours but, to do this at scale, you kind of have to use AI because there’s millions and millions of data points to tap into.
Once we understand that, we also need to figure out how to analyse the outputs of that data so we can optimise the different messages and say the right things to the right people to move them through the funnel.
This could really solve some of the measurement issues that PR has had over the years. It feels like machine learning might be something you’d use to measure results over a long period of time rather than a short campaign, right?
Yeah, definitely.
We’re ultimately getting to the point of automation of process ease. From a PR standpoint, those data points are being leveraged from different technologies, and the opportunity lies in infusing that into the tools we use every day. So, something that used to take teams days or weeks to do is now built into your software and can be done immediately.
I don’t think we’re quite there just yet with full automation, but for about 6–7 years now there have been tons of tools that can do incredibly powerful things throughout the PR workflow.
You said that you did some research to look at how AI was being used across the Edelman business. Do you find that the technology used in the workflow varies in different markets?
Yeah, we’ll see the same technologies and platforms being leveraged in different ways. Some of the technologies I brought into Edelman and scaled them to a global level, so I’ve ultimately been trying to find the different ways we’ve been leveraging these data points and then educating everyone on how to use them.
It’s pretty cool, because you always find something new. Our teams are smart teams who have been focusing on AI for a long time, so it’s always good to see how we’re using the technologies and the results they’re getting.
How do you set about doing a research project like that?
Carefully! It’s a lot of emails and trying to find out what’s out there, but also at the same time having a remit on the technologies and knowing who has access to what.
It’s amazing because of the depth and breadth of the technology Edelman uses in the marketing and communication landscape, so we need to find out how to stitch together the different ways we’re using AI and then use that intelligence to create solutions for other clients in different markets.
Basically, it’s a lot of education and a lot of hunting around.
So for people who might not be setting out to do global research, how do you recommend they unearth the tools that are being used even just within a team?
Definitely have someone dedicated to it.
You should hopefully be able to find out what contracts have been signed and, if they are free or freemium, even just sending out an email to your database can bring back some interesting results. People who are at the forefront of leveraging technology are usually more than happy and excited to talk about how they’re using it which tends to help.
The people who are going through this research process might realise there are some gaps in their workflow, what advice would you give to someone who finds out they’re not using technology enough?
I would just say that budgetary limitations tend to be the biggest blocker.
Take one area of PR, like media relations. I would try and find out the biggest technologies in that area and have a conversation with them to discuss where they’re thinking of leveraging AI.
They probably already have hundreds of thousands of clients that are providing this information and feedback on a regular basis and they’re probably already exploring how to bring AI into their technologies. You’ll know so much more from talking to technology vendors.
Then find a pilot opportunity, prove it out, and see whether the benefits outweigh the costs, and you can then push it out from there. That’s the true innovation cycle. There’s always going to be something new, but you want to hitch a ride just to the 2–3 best out there.
Because of how many different clients we have at Edelman, we have tons of different technologies that we leverage across the board, but we also have to understand the technologies our clients are already using to determine if there’s a way we can leverage those successfully too.
Do you review the benefits of the technologies you use — or don’t use, for that matter — regularly?
Yes, especially when we’re exploring it in the first instance and especially for processes that take a lot of time.
For example, something that used to take 20 hours but with a new technology that costs a fair amount of money, you can cut down to two hours — that’s a huge benefit. You have to do that math because, even though there’s a cost associated with it and the client will be paying that money, the technology might be a better place to spend that money if it’s giving you the time to do other things.
This whole concept of manpower vs trying to introduce technology to reduce manpower which can then be used on creative or better analysis seemed brand new back in the day. Now, with AI and technology we can move away from that manpower and get more value out of our time.
I guess before people get overwhelmed, the key is to start at home and look at the technologies you already have and are using and go from there, right?
Yes, especially from a media perspective.
You probably know all of the top companies and blogs in that sector, so you can then see who’s talking about it and see which companies resonates most with you.
But yes, starting at home and seeing if there are teams that are already leveraging technologies is really useful, particularly when you drill down into how they’re using them. Sure, the technologies might not have specific AI functionalities, but they’re showing progress towards an automated future.
Looking to the future, are chatbots an area that PRs should be looking at? Or, perhaps more importantly, to what extent do you think PR will introduce chatbots into campaigns?
It really depends on the company and where their communications sit in relation to their customer service as opposed to the social and marketing side of things. Chatbots can sit in multiple different places right now.
The important thing is to think about how you can get the customer what they want. With a chatbot people are usually there for a specific need, they’re not specifically there to get entertained so you have to think about what the customer is going to get out of it if you’re using it in a PR context.
Don’t try to reinvent the wheel. Instead, you can learn from what’s already been done and take what’s most relevant for your business and expand from there.
Finally, where do you find your inspiration? Is there anything you read regularly?
I think some people might remember Google Reader — I was an avid Google Reader fan.
I use Feedly now, so I still read a lot of RSS feeds. I probably scan about 1000 article headlines per day, which might be on Twitter or somewhere else, then I skim another 20–30, and then deep dive into another 5–10 per day.
I read a lot of the top tech publications. There are plenty out there, especially if you look at the big companies that focus on AI, or even venture capital sites, like CB Insights. There’s no shortage of content, but once you start reading certain sites, you’ll quickly be able to see if it’s relevant for your business within about a week or so.
You can listen to the full interview with Adam on the PR Resolution Podcast here.
