We interview AI ‘wonder geek’ Alex Clelland, Director at Houston

Few professionals in the communications industry combine a deep understanding of strategy with a genuine enthusiasm for technologies like AI quite like Alex Clelland, Director at integrated communications consultancy Houston. Naturally, we might call him a wonder geek!

Alex’s 20 years of comms experience includes in-house roles at Macquarie, RBC Investors and Fitch Ratings, alongside senior roles at well-respected agencies where he has witnessed the promises and pitfalls of tech evolution in comms. He shares his journey from scepticism to advocacy for AI, recounting how tools like ChatGPT transformed his approach to audience insights and data-driven decision-making and shares his thoughts on the importance of training and asking awkward questions.

1. Can you give us an overview of your comms experience and when it was that you started to take an interest in AI?

I have been in comms for over 20 years now, evenly split between in-house in financial services and working for agencies across a variety of sectors. I have spent a lot of that time wrangling with technology that overpromises and under-delivers.

When the hype around ChatGPT and other Large Language Models (LLM) exploded, I was intrigued but sceptical. Suppliers had been dropping terms such as machine learning and artificial intelligence into conversations for a while without any discernible improvements.

I did some digging and started playing around with ChatGPT and was impressed. I did have some arguments with it (I like to treat it as an intelligent colleague) but it got better. The constant upgrades have meant its potential has increased massively over the past two years.

2. We know you're a bit of a whizz when it comes to AI and I understand you’ve put together some effective tools to support you with uncovering different audience perspectives of your clients’ reputations. Can you tell us more?

The comms industry is not known for its over-familiarity with maths and statistics, despite the vast quantity of spurious polls it produces, and I am no different in that respect. But when I saw how AI could transform data management, I decided to take the plunge and studied data for business management at the LSE for a brief period. It transformed how I saw our business.

If you ask comms people what data they have available to them that will improve their work, most will not be aware of even a small proportion of it. The truth is that when we combine the data from our suppliers and our client work with the data that our clients hold and are willing to share, then we have a veritable goldmine. The challenge is that the industry has very few miners (despite our working conditions being quite a lot better than those of the mining sector).

We are fortunate now that Houston is part of Brand Potential, in that data scientists are more easily found in the brand and marketing world than in comms, so we can draw on their capabilities and combine them with our comms expertise to generate insights and actionable intelligence.

This can range from the relatively straightforward issues around measurement, such as using analytics to show correlation and causality between activity and outcomes, to identifying hostile actors online and mapping their network and influence to inform a response, and whether one is even warranted.

Personally, I prefer the more investigative side of things. For too long, the industry has been in the business of telling clients what they already know and what to do about it. You get a different reaction when you tell clients what they don’t know and why they need to do something about it.

3. It feels like you're a fairly early adopter when it comes to taking a deep-dive into using AI tools in the workplace. How do you go about training your team at Houston to use AI effectively?

Training is essential to making any progress in this area. If you let people go off half-cocked, it can backfire in multiple ways. This can range from unwittingly breaching confidentiality by plugging client data into a LLM or being frustrated about the results you are getting because you aren’t asking the right questions. A little knowledge can be a dangerous thing.

We started training people very early on and educating them on what AI meant for comms and what it meant for us specifically, in terms of our roadmap. When it comes to training people on different tools, we are careful to focus on those who need the tools and training for their work. Who needs what and why? Training on ChatGPT or Copilot is for everyone, but overloading everybody with information and training on more specific tools and techniques can be counterproductive.

4. Are you facing any challenges with AI right now?

Yes! I don’t think I will ever stop facing challenges with it. I am a nightmare with suppliers because if I can see how their tech can solve a problem and they haven’t developed that functionality yet, I want to know why. I am not deliberately being a pain, I am just curious to know if there is something I don’t know (and that is a lot).

5. It's unusual for someone so senior to be so geeky when it comes to AI! Where did the interest come from?

It springs from several sources. One is a constant frustration with the standard of tech capabilities in comms over my career, in comparison with adjacent sectors such as marketing. Another is my deep-seated insecurity that there is something important happening in the world that I don’t yet know about. It’s a manifestation of a broader hypervigilance, but we won’t go into that.

Finally, a love of science fiction. This is probably key. I read the Hitchhiker’s Guide To The Galaxy by Douglas Adams when I was 11 and imagined a handheld device that could tell you anything you wanted to know about the universe, even if a lot of it was ill-informed, biased or incorrect. Now I have at least three such devices within reach for most of my day.

6. What advice would you give to your peer group about AI and using it at work?

Don’t believe the hype. Ask awkward questions. Try and break things. Do tell your employer what you are doing, why you are doing it and what you find out. Resist the temptation to get tools to create or do things that might lead to your imprisonment here or in other countries.

7. How are your most junior colleagues within Houston using AI compared to how you use AI?

The most notable difference is that they were more advanced in using it in their personal lives than at work, whereas I was the other way around. Some of them were doing very creative things with it from planning holidays and activities in other countries, to creating treasure hunts for their friend over the weekend. And there I was, trying to correlate press releases with website traffic. So sad.

8. How do you see AI really helping comms professionals over the next few years?

It’s like any technology – if you put rubbish in, you will get rubbish out. We need to start thinking a bit differently, in line with my comment above about what we don’t know. What don’t we know and how can we find out? Why don’t we know? What can we do about it? I don’t want to get into known unknowns and unknown unknowns really, because I rapidly end up in an existential crisis.

9. What are your favourite AI tools a) in the workplace and b) outside of work?

For general use, I default to ChatGPT because I have invested so much time in it and we have a very close relationship now, both personally and professionally. I ask it questions and it reminds me of things I have previously said or asked about, and how this needs to be taken into account. I like it when technology makes me feel inadequate.

I am also a huge fan of Open Source Intelligence (OSINT) tools for both personal and professional reasons, all of which integrate AI.

Wayback Machine has been archiving the internet since 1996, so you can see what changes have been made to a website and when, which can be very useful for all kinds of reasons in issues and crisis management or to force companies to honour offers and deals in your personal life.

Tools such as ZoomEyeGPT let you search for assets in cyberspace without having to learn all the complex coding etc. It currently offers examples including “publicly available security cameras in Russia" and "Starlink devices in Ukraine", which reflects the interests of those using it at present and is definitively NOT what I use it for.

Then you have sites such as NotEvil which enables you to search the Dark Web while filtering out the unsavoury content, which can be useful for research, but most useful and fun are the cyber investigation platforms such as Maltego, which can help you find out all kinds of interesting things about people.

Thank you Alex!

For a free download of our full Annual Salary Guide 2024, click here.

Time for a change? Check out our latest jobs!