How To Keep The Human Factor In Learning And Development With Andy Candler

You have the claim of creating the first digital agency in 1987. It would be great to go back there to the beginning and how you got into L&D.

So it’s a funny story, and I’ve just written a book about it recently. We were selling presentation software. I left school in 1986 at 16, and in 1987 I met someone and we started working together selling presentation software. One of the companies he met with was Amstrad, Alan Sugar’s company, and they weren’t interested in buying the software. All they wanted was to buy a presentation from him.

We struggled because we were selling the software, but we didn’t know how to use it properly ourselves. I was looking at some of the old files the other day and some of them were timestamped at three in the morning. Eventually we produced a rolling demonstration system that ran on Amstrad computers in shops. Effectively, we turned that computer with a big black screen into an advertising space.

In the early 1990s we were working with a lot of UK retailers, many of which don’t exist anymore, and they said to us, “You’re really good at producing things to help our customers buy products. Can you start helping our store colleagues sell products?”

That’s when we started producing digital learning. At the time it was delivered on floppy disk, then CD-ROM, then DVD-ROM. In 1999 we produced e-learning for British Telecom, and all of it was hand-coded because there were no learning development tools or software platforms available at the time.

We also created one of the first cloud-based LMS systems in the world in 1999 to support BT’s partners, wholesale teams, indirect channels, and internal staff. That’s really where I entered the learning space.

From there my business continued operating across two channels. One focused on interactive digital systems and signage in retail environments, while the other focused on digital learning. Over time we stopped hand-coding and started using content development tools because they became much more capable.

I left agency life two years ago and decided to become independent and freelance. I had to choose which side of the business I wanted to focus on, and I chose learning and development because that’s where my passion really sits.

Now I deliver everything I used to deliver within an agency structure, but independently. I have a team in India that produces content for me, and I’ve got a huge network of specialists. So if a client comes to me and says they want specialist learning I don’t support, I’ll reach out to experts I trust and collaborate with them. For me, it’s very much about collaboration. I don’t do everything myself.

Digital learning feels like it changed massively after the pandemic. Before that it existed, but since then there’s an argument that it became oversaturated and sometimes used as a default solution. 

What do you think digital learning looks like now, and where do you think it’s most relevant today?

The pandemic suddenly gave organisations permission to deliver learning online because they had no other option. Before that there were still a lot of companies saying, “No, this has to happen in a classroom.” Then the pandemic happened and everyone stepped back and said, “Okay, let’s try this online.”

The problem was that many people tried to replicate the face-to-face classroom experience digitally, and that simply doesn’t work.

I’m a huge believer in blended learning. For my clients I deliver digital learning as part of the solution, but it’s not always the solution by itself. Often classroom learning still needs to exist. Sometimes a podcast, webinar, infographic, or white paper can be the best way to communicate something effectively. A lot of organisations did suddenly shift online during the pandemic, but I don’t think many of those experiences were particularly good. It was necessary at the time.

Now I’d say things have settled down again. In many ways, organisations have reverted back toward traditional approaches. What’s interesting is that some facilitators now run face-to-face sessions while also allowing people to dial in remotely because there’s now an expectation of flexibility.

The major challenge for learning experiences now is definitely generative AI. During the pandemic businesses were producing poor training because they had to move quickly online. Now many people are producing poor training because they think generative AI is the complete answer, and it simply isn’t. It’s just another poor learning experience if it’s used incorrectly.

AI is a common thread that comes up in these conversations. There’s a strong feeling that AI won’t replace genuine learning and development because it lacks something fundamentally human. 

What’s your perspective on that?

AI is a tool. I remember when spreadsheets first appeared in the late 1980s. Before that, people were doing calculations manually on paper. Suddenly you had Excel and Lotus 1-2-3, and it completely transformed the way people worked.

I think generative AI is at a similar stage right now. People think it can do everything for them and replace decades of expertise from experienced consultants, but it can’t. It delivers a poorer experience.

I often describe generative AI as being in its teenage years. Two years ago it behaved more like a five or six-year-old. It’s developing quickly, but I still think it will take years before it offers truly high-value outputs independently.

At the moment, all it really does is scrape and remix what already exists. There’s no originality. It’s like a cover band rather than the original artist.

That said, I absolutely use it as a tool. When I’m producing learning content, I still use human beings to write the material. But I use large language models for things like grammar checks, spell-checking, or reviewing whether a piece of content aligns with the intended learning objectives.

For example, I might feed in a human-written script and prompt the AI with something like:

“This learning that we’re creating has been designed to achieve these goals. Do you think this section achieves those goals?”

That’s a different use case from asking AI to generate the entire learning experience from scratch. It becomes more of a review and support mechanism rather than the creator.

We also use it for repetitive and mechanical tasks. For example, if we have a thirty-page script, we might prompt it with: “Here’s a thirty-page learning script. Create a two-paragraph synopsis.”

We would still edit that synopsis manually because it’s never completely accurate, but it saves a significant amount of time compared to doing the whole thing ourselves. So for me, AI is extremely valuable for efficiency work, repetitive tasks, summarisation, checking, and supporting processes. It’s excellent for the mundane and mechanical side of production, but it’s not the complete answer when it comes to creating meaningful learning experiences.”

You mentioned prompting and giving AI detailed context. Have you found that to be one of the most effective ways to use it productively?

Yes, absolutely, but only if you prompt it properly. You need to give it proper context. You have to tell it who the audience is, how the audience will access the learning, why the learning exists, who the client is, and what specific problem you’re trying to solve.

A proper prompt can easily be half to a full A4 page before you even provide the actual learning material you want feedback on. You can’t simply type, “What do you think of this?” and expect meaningful output. A lot of people still use AI like a search engine, and that’s one of the reasons they get poor-quality results.

The prompts need structure and intent behind them. For example, when reviewing learning content, the prompt might include:

• Who the learners are

• Their existing level of knowledge

• How they consume the learning

• What business outcome the learning is trying to achieve

• What behaviours should change afterward

• What tone the content should have

• What success looks like

Only once all of that context is established do you then provide the script or content for review.

I know somebody producing learning content whose prompt document is six A4 pages long. They’ve invested a huge amount of effort into shaping the context and defining exactly what they want the AI to do. Because of that, the quality of the output is significantly better than what most people get.

That’s really the key lesson with AI right now. The quality of the outcome is directly connected to the quality of the prompting. The more thought, structure, and context you provide, the more useful the tool becomes.

So yes, it’s a fantastic tool for repetitive and mechanical work, but it still requires human thinking, human judgement, and human experience to guide it properly.

You describe yourself as a plain-speaking learning consultant. I imagine a lot of that involves demystifying L&D jargon and helping businesses understand ROI more clearly. 

What best practices have you seen for tying L&D initiatives to measurable return on investment?

The reason I describe myself as a plain-speaking learning expert is exactly because the L&D industry is full of acronyms, frameworks, and terminology that can become overwhelming and difficult to navigate.

I tend to tell things exactly as they are, and that’s what my clients appreciate. My website even includes a glossary of terms built with Claude, not just explaining definitions, but explaining how those concepts actually impact businesses in practical terms.

I’ve been fortunate because a lot of the learning I’ve produced has been directly tied to retail sales and customer behaviour, which makes ROI easier to measure.

For example, with Sonos, they wanted retail staff across the world to understand a new speaker product. We delivered the learning in eleven languages. But instead of focusing on technical specifications, drivers, or impedance, we focused on what it feels like to own the product and how it improves your life.

We used day-in-the-life storytelling and emotional positioning. The same applied when we worked with Google Nest products around security, heating, and lighting.

The learning wasn’t just about product knowledge. It involved helping retail staff emotionally connect with the brand. If someone walked into a retailer looking for a speaker and had ten brands to choose from, we wanted the sales associate to genuinely feel connected to Sonos and naturally recommend it.

We saw the same thing in travel. We created learning experiences that made travel agents more likely to sell a particular skiing holiday or premium package over competitors.

In those environments ROI becomes straightforward to measure. One travel company spent £70,000 with us and generated roughly £700,000 in additional revenue. Travel agencies that previously didn’t sell skiing holidays started selling them, while agencies already selling them began upselling premium packages. The travel agency’s revenue increased by around 22%.

Internal business transformation projects are more difficult because the outcomes are less directly measurable. One client solved that by creating two business cases, one showing the benefits if training happened, and another showing the cost if it didn’t happen.

There’s also another useful metric beyond ROI, which is ROE, return on engagement. If you deliver non-mandatory learning and engagement levels are exceptionally high, that in itself can be a meaningful indicator of success.

Ultimately, if you can clearly tie learning to a business problem and demonstrate that the problem has been solved, then you’ve successfully measured impact.

It sounds like retail environments make ROI easier to measure than internal business transformation. Are there other measurements that surprised you when working with retailers?

Definitely. People naturally assume the measurement is simply how many TVs or microwaves were sold after training. But one retailer taught me that another critical metric was product returns.

If a customer buys a television and then discovers they can’t connect their existing devices because the salesperson didn’t properly explain compatibility or adapters, the product often gets returned. So one of the key success metrics became not just how much product left the store, but how little came back through returns. That was a really interesting shift in perspective.

Quantitative data is obviously important, but qualitative stories and feedback matter too. How do you balance those two sides when measuring learning success?

It completely depends on the problem you’re trying to solve. I run discovery workshops with clients where they come to me with a challenge they want to fix. At the very beginning of the session, I ask them a simple question: ‘How are we going to measure this project as a success six months from now?’

Quite often they haven’t actually thought about that. They just know there’s a problem.

Sometimes they’ll identify metrics immediately, and sometimes there’s a lot of head scratching. I then ask whether they actually have reliable source data available to measure those outcomes properly.

My philosophy is always learner first. We begin by understanding who the learners are, where they are, how they like to learn, when they like to learn, and what environment suits them best.

Then we revisit the business problem. Sometimes the problem they identified initially is correct, but quite often, after discussion and reflection, we discover that the real issue is something slightly different. From there we explore potential solutions, which may or may not involve technology. I always say digital learning is never the full answer by itself, learning must bea blend of solutions. 

At the end of the workshop, I ask the success-measurement question again. Usually by that stage the client has a much more mature understanding of what success actually looks like and how it should be measured.

Then, depending on the project, we measure outcomes after one month, three months, six months, or nine months. If something isn’t working, we feed those findings back into the process and revise the learning accordingly.

What advice would you give to companies who are researching LMS technology to use for their programmes?

I often begin by asking clients whether they even need an LMS at all. The best learning is learning people can access within seconds wherever and whenever they need it. If organisations aren’t training their staff, employees will simply go elsewhere to learn whether that be YouTube, TikTok, or now generative AI.

So sometimes the answer isn’t an LMS. Sometimes it’s simply curating high-quality learning experiences and making them easily accessible through something as simple as a WordPress site.

Of course, there are industries where compliance and regulation make tracking essential, and in those cases an LMS absolutely matters. But I always say the platform should come last. Solve the business problem first, then choose the technology.

I’m platform-agnostic. Most LMS systems fundamentally do the same thing, even if vendors claim otherwise. What really matters is the company behind the LMS.

If you’re entering a three-year relationship with a provider, make sure they’re going to support you properly for those three years. Some companies only care about selling licences, while others genuinely listen to customers and continuously improve their product.

I used to recommend an LMS that had a beautiful interface and fantastic pricing. Then the company got acquired, the people who cared disappeared, and the product stagnated. I can’t recommend it anymore because the company behind it lost its identity. So my advice is always pay close attention to the people behind the product and how they respond to you during the sales process. That matters enormously.

You recently became a business partner with Kim Ellis at L&D Free Spirits. How did that relationship develop, and what appealed to you about the community?

I was made redundant in July 2024 after expecting it for about six months. Before that I’d spent 34 years running my first agency, which grew to 26 staff and around £1.5 million turnover. Then I spent several years helping another agency transition from a lifestyle business into a growth agency before it was acquired.

After leaving, I decided I wanted to work independently. I didn’t want to build another large agency. I simply wanted to work for myself and enjoy the freedom that comes with that.

But suddenly I found myself wondering how to structure this new life. Should I be a sole trader? A limited company? VAT registered? I was effectively starting from scratch again.

At exactly that point I met Kim, who told me she was building L&D Free Spirits. The timing was perfect. I quickly realised the community wasn’t focused on teaching people how to do their jobs. The assumption was that members already knew how to deliver exceptional work. Instead, the community focused on helping people build sustainable businesses around their expertise.

That was exactly what I needed. I started attending all the sessions, especially the Friday Fiestas, which genuinely became the highlight of my week. I like consistency and structure, and what I missed most about agency life was the people.

The Friday Fiesta became my version of a business stand-up meeting. It just happened to involve people from different businesses rather than colleagues in the same office. We talk about challenges, successes, failures, travel, television, language, and life in general. There’s a real camaraderie to it.

Kim then invited me to start delivering webinars, which was initially a little intimidating, but I really enjoyed it. Over time I became more involved in supporting the community and mentoring others.

Eventually she invited me to become part of the business officially. I didn’t immediately say yes because I wanted to step back and ask myself whether I could genuinely contribute at the level expected. Then I realised I wasn’t actually being asked to become someone different. I was already doing the work naturally within the community. It was simply becoming official.

I genuinely believe in the community. It feels like my tribe, and I get just as much from supporting others as I give back.

You’ve spent a lot of time travelling, especially in Colombia. Has that experience changed the way you think about learning, work, or life more generally?

Before 2018 I had what most people would probably describe as a successful life. I was married, running a successful business, travelling extensively, living luxuriously and owning lots of expensive things.

But I wasn’t happy. In 2018 I decided to completely change my life. I got divorced, got rid of a lot of the material things, and realised life was much more about experiences than possessions.

Around that time I met my current partner online, who lives in Bogotá, Colombia. We’ve now been together for five years. Because of her circumstances we currently live between countries, but eventually we plan to live together permanently.

I’ve been to Bogotá maybe five or six times now, and spending time there has been incredibly grounding. Colombia is still a developing nation, and many people genuinely live hand to mouth. What you earn that day is what you spend that day. If you don’t earn money, you don’t eat.

Experiencing that reality up close changes your perspective very quickly. It’s made me appreciate how fortunate we are, and it’s also inspired me to try and build stronger business relationships between Colombia and the UK. I’m now involved with the British Colombian Chamber of Commerce and trying to find ways to use my experience to support Colombian businesses and people if I can.

I’m still figuring out exactly how I can contribute meaningfully, but I’m actively building those relationships.

It’s also been personally transformative. I’ve been learning Spanish for eight years now, so there’s been a huge cultural and linguistic learning curve as well. Overall, I think it’s made me a better person and given me a much richer life perspective.

It’s also one of the major reasons I’ve remained freelance. I value the freedom to work from Colombia, the UK, or anywhere else. Flexibility and freedom matter much more to me now.

If there were one or two things you could change about the L&D industry right now, what would they be?

I’d like people to value L&D professionals more, especially freelancers. Going back to generative AI, I’d like organisations to understand that AI is useful as a tool for repetitive tasks, but not as a replacement for genuine expertise and creativity.

Freelancing is a rollercoaster. Some months are brilliant and some months are difficult. You genuinely sit there at the end of the month wondering whether you’ve got enough money to pay yourself properly. I’d love to see businesses place more value on what experienced L&D professionals bring rather than assuming ChatGPT can create a meaningful learning experience on its own.

Beyond that though, it’s a wonderful industry. One thing I’ve noticed about the L&D world is that people are genuinely supportive and generous with one another. There’s no jealousy or competitiveness compared to some other industries.

In communities like Free Spirits, if somebody lands a huge contract, everyone celebrates it sincerely. Nobody sits there resentfully wishing it had been them instead.

I think most people in L&D genuinely want to help people improve, whether that’s helping a business grow or helping an individual develop. Yes, we get paid for it, but I don’t think many people enter this industry purely to make money. There’s a genuine sense of purpose behind it, and that’s what makes it such a rewarding community to be part of.