How AI is Powering Knowledge Transfer in Today's Workplace with Donald H. Taylor

There was a lot of talk during the pandemic and lockdown that conferences would go. They've come back and there's

would go. They've come back and there's a reason for that which is to do with trust and to do with the quality of information transfer you get face to face. Any major takeaways from from that

face. Any major takeaways from from that event? Lots of possibilities on the

event? Lots of possibilities on the horizon as well still as the fear and the uncertainty, but at least there's some sense of knowing where we can go.

[Music] Hi everyone, welcome to the business AI playbook. My name is Dominic. I'm the

playbook. My name is Dominic. I'm the

CEO and founder of Colossian. We focus

on real world examples. And today I have in our studio Donald Taylor who is chair of the learning technologies conference.

uh have been running the learning sentiment survey uh for over a decade and uh has had multiple engagements over 30 countries as a international speaker

and is really excited about the topic we'll discuss today. He's been

consulting various organizations and experts worldwide. All right. So Don,

experts worldwide. All right. So Don,

thank you so much for coming today. It's

so lovely to have you here as a guest and uh you've been uh uh you've been having like quite a busy week last week because you organized the learning

technologies conference in London 2025.

Uh any major takeaways from from that event? It's very important to plan your

event? It's very important to plan your recovery period. This time I took my

recovery period. This time I took my Friday off. I had a couple of meetings.

Friday off. I had a couple of meetings.

I've discovered that's the most important thing. But actually seriously

important thing. But actually seriously whereas last year 2024 it was a time of what I call AI shock. Everyone was talking about AI but

shock. Everyone was talking about AI but didn't know what to do with it. This

year 2025 a whole different approach.

Lots of very concrete things that we can do with AI. Lots of possibilities on the horizon as well still as the fear and the uncertainty. But at least there's

the uncertainty. But at least there's some sense of knowing where we can go.

Yeah. Absolutely. like uh it's a it's a whole adoption cycle and we see it uh real time happening right but it's also quite promising that uh you know this

conference is primary after the pandemic uh you know like your numbers are not dropping but there is a steady uh increase even possibly in terms of

interest and uh and vendors possibly like ourselves and also um industry players are showcasing and sharing their knowledge more and more around these events. Actually, we'll come back to

events. Actually, we'll come back to this later on during our conversation, I'm sure, the importance of face-to-face conversation as part of knowledge transfer. But yes, I there was a lot of

transfer. But yes, I there was a lot of talk during the pandemic and lockdown that conferences would go. They've come

back and there's a reason for that which is to do with trust and to do with the quality of information transfer you get face to face. Absolutely. Um it's a it's

an important psychological aspect. Um

and uh the reason why we are here today as well is because uh uh I think it's it's one of my uh favorite topics personally because I built Colossian our company with a mission to make knowledge

transfer easy. Before this conversation,

transfer easy. Before this conversation, I mentioned to you that I believe we are doing so many things still unwell in terms of knowledge sharing within our own organization. And when we had the

own organization. And when we had the briefing for this episode, you mentioned that this is one of the most uh important challenges to solve uh using traditional method methods and also

using AI. Um can you elaborate on what

using AI. Um can you elaborate on what you've seen over your career so far? why

knowledge transfer uh knowledge retention of a specific organization is so important. What trends are you

so important. What trends are you seeing? What uh what's emerging now in

seeing? What uh what's emerging now in the field? Yeah, I I can't answer all of

the field? Yeah, I I can't answer all of that in one short answer, but what we'll do is we'll explore these themes over the course of the conversation, but I think the key thing that I'd look at is

how important and pressing the need for understanding what you know in your organization is now in comparison with how it was when I was a kid in the

1970s. When I was growing up in the

1970s. When I was growing up in the 1970s, uh, Oantomo do this, um, study of the value of the intangible assets of an organization. So they show that the

organization. So they show that the standard and pores 500 companies, their balance sheet was roughly 20% intangible

and 80% tangible assets in the 1970s.

that switched around and it is now more like 15% tangible and the rest 85% intangible

assets. What does that mean? It means

assets. What does that mean? It means

that the value of companies is embedded in stuff that you can't pin down and sell. Some of that is knowhow. Some of

sell. Some of that is knowhow. Some of

it's algorithms and methods. A lot of it is people. Now people are not like

is people. Now people are not like machinery. They can get up and walk. And

machinery. They can get up and walk. And

if so much of the value of a company is embedded in the people, then you've got to make sure that you're looking after the people. You also have to make sure

the people. You also have to make sure that when people walk out of the company, as they do for various reasons, they don't take with them the secret source of the organization. That

knowledge, that knowhow about what and how to do things doesn't disappear with them. And I think for me, that change is

them. And I think for me, that change is the crucial change. And there's a lot of other stuff as well, but that's the crucial change that's taken place in my lifetime. And do you see an emergence of

lifetime. And do you see an emergence of uh of this challenge like uh becoming harder because uh trends for example showcase that uh employee retention is

is is is going down. Um loyalty towards the employer is not the same as 30 40 years ago. Yeah. I mean loyalty comes in

years ago. Yeah. I mean loyalty comes in different packages. I mean sometimes

different packages. I mean sometimes loyalty is just practical. I in the old days particularly in the UK you had pension schemes which meant if you stayed with an employer for a long

period of time it was worthwhile sticking around and then you left. Um

that loyalty wasn't based on love. It

was a very practical view of your of your future. Um but those pension

your future. Um but those pension schemes don't exist anymore. And it is also true that I think the phrase is work won't love you back. I think people

have come to understand that it's a transactional relationship or more of a transactional relationship than they believed in the past between themselves and their employer and so they don't have to stick around. There's less of an

emotional reason to do that and then if they leave well yeah they can leave. So

now do does it mean that that people are staying in work for shorter periods of time? I'm not sure. I think the v the

time? I'm not sure. I think the v the data on that is varied, but they certainly they certainly don't have to stick around in the way they used to have to and they are probably more able

to move particularly if they're a high value individual more able to move to another job in a parallel or completely different sector because they've got

transferable skills.

One sector which we'll come back to look at as an example of that in the course of this conversation is the rail sector in the UK which sounds a bit odd but it's quite a good example of some of the

issues that exist in keeping hold of skilled people and long-term effects of failing to do so. Um, and in the rail

sector, if you're a driver, you're unlikely to leave because there aren't many jobs for trained drivers outside the train sector, obviously. But if

you're a project manager, there are lots of jobs for project managers all over.

And your skills are very transferable.

And so there's definitely it's not the case that everybody is staying in jobs for shorter periods of time but some people are much more likely to transfer because they can

get a better job elsewhere. Yeah. Uh I

see I see the challenges there. And uh

how do you see how much of a focus area is this for relevant leaders professionals? uh how how much are they

professionals? uh how how much are they focusing on on solving this challenge or is this more like um so sort of a uh not

prioritized so far in the industry I I I don't think it is sufficiently realized except in specific industries where it's it's absolutely seen as being crucial so in rail is one in the UK and the rail

industry they are absolutely aware there is a generational cataclysm coming which they've seen coming for probably about 10 years but have been unable to do

anything about. The same is true for oil

anything about. The same is true for oil and gas which is different. Rail will

continue as rail. It will change. Oil

and gas is moving towards green energy.

You still need people to handle legacy systems and you need some of those engineers to transfer their skills to work with new technologies from oil to electricity. But some of the fundamental

electricity. But some of the fundamental things are the same. In those two industries in particular, it's seen as being um a real issue. I don't think

it's seen as being so much of an issue elsewhere, but it probably should be and one of the reasons is demographic. You

know that we have a across Europe a low birth rate. Uh it's insufficient

birth rate. Uh it's insufficient typically 1.6 1.5 uh birth rate whereas you to replace your population needs to be around 2.2 to 2.3. And this means

that we are not having enough people coming into work to maintain the levels of employment we've had in the past.

Which means that over time we are certainly going to have under supply of qualified people.

Meaning that again, if you're a highly qualified or highly valuable person, you're likely to move on elsewhere. I

don't think enough people are taking enough notice of that. And I think for learning and development professionals, this represents an opportunity. If you're in the business

opportunity. If you're in the business of L & D, you can ask yourself, am I in the business of creating courses or I'm in am I in the business of creating a

strategic outlook for my company to assist it in the future? And if the f if most organizations are going to suffer in the future by not being able to get a hold of enough qualified people, they

want to hold on to them and what they know. If L & D focuses there, it can be

know. If L & D focuses there, it can be real strategic, but I don't think in L &D people are thinking about that at all.

Yeah, I say I see that as well. Like

that's that's why I as about the sort of prioritization area of this. Um, and you know, my my tricky question is to sort of get AI into into this uh into this

topic is let's say tomorrow a CEO tells an HR or R&D leader that we are losing our knowledge. we're using uh basically

our knowledge. we're using uh basically the uh this key uh information fast.

Let's throw AI at the problem. Where

would you start or is that even a a valid statement like is AI the uh the best solution we have for this problem today? I think it's unlikely a CEO would

today? I think it's unlikely a CEO would put it in terms of we're losing knowledge, but they would likely say we can't do what we used to be able to do.

So they can see the performance impact.

I think it is likely they'd say, "Let's throw some AI at it because AI is so hot." And I do think it is a legitimate

hot." And I do think it is a legitimate thing. AI is both a cause and a savior

thing. AI is both a cause and a savior in this situation. So AI enables the spread of information, makes it much more likely that highly qualified people can get work elsewhere. It also enables

us to identify the tacit and implicit knowledge that we have in the organization and enables us to share it,

codify it much more uh effectively.

There are there there are so many ways that can be done. I know that you've got some ideas yourself. Yes, exactly. So I

I done some research and uh one of the areas where we've been uh sort of uh heavily invested into is finding buried knowledge. So basically with

knowledge. So basically with enterprisewide search systems uh called rags um uh now like basically trying to trying to find insights and also

correlation within the hidden information the data links of of a company. How how does that work? Can you

company. How how does that work? Can you

tell me more about it? It's this is the sort of thing makes me super excited.

What are you doing? So basically like uh you connect all the relevant knowledge bases and and databases to an AI system of course like enterprise compliant with the right security measures and uh you

try to uncover insight. So basically if you um if you don't know about how to on board to a specific uh uh software or

you don't know uh a specific uh steps for uh registering at a tool uh based on previous conversations that there's hope

in this whole thing that hopefully the AI system recognizes that it's not outdated information will showcase it to you and deliver it to you in instead of

you having to search for manuals. Um I

think this is quite impactful for this topic that we are talking about. Have

you seen this uh in practice eventually this enterprisewide search systems based on unstructured data? Yes. Uh and there are a number of them out there and I was I've been introduced to quite a few

recently. I won't mention the names

recently. I won't mention the names right now, but they tend to have the same approach, which is either they're looking across the data lakes, the the

the repositories of formalized or semiformalized information in terms of documents and what have you, or they're looking at conversations typically in Slack channels and other areas where people are talking at email and they're

surfacing and and codifying the knowledge in that way. They all are aiming to do the same thing which is to make it easier to find the information

which is there but which is spread very thinly across a variety of sources.

Absolutely. Uh and um um from from the perspective of our conversation I I see how AI could help with this. For

example, how about like hidden expertise where for example um the expertise is is indeed spread out and um based on what I

saw you have to use like certain knowledge graphs and and mapping uh methodologies to rebuild that expertise that's like uh lost in a way. Um have

you seen uh practices uh doing that or or successful implementation? Yeah,

sure. So I mean at Learning Technologies we had Yep Hansgard speaking. Last year

we had um Stalin Hunter III speaking. Um

Yep is from Denmark. Uh and uh Stalin is actually from the US but he's based in Sweden. And it seems that the Nordic

Sweden. And it seems that the Nordic countries have a sort of uh an area of expertise thinking about this. I don't

know why that is. Um and we also had of course uh keynoting this year. We had

Daniel Hume speaking talking about he's the um chief AI officer for WPP and he was talking about uh his company Satala that he runs and how they do

organizational analysis across that and also across the uh WP sorry across

WPP. Um and what they do is they find

WPP. Um and what they do is they find not what the organizational chart says people have relationships to each other which is the formal structure but the

relationships the people actually have in terms of their electronic communications and finding who are the nodes. So who are the who are the points

nodes. So who are the who are the points that are most connected who are the bridges between different groups and so on. Once you have this information,

on. Once you have this information, uh it is possible much more rapidly to identify areas of expertise that aren't selfidentified. Now I think you have

selfidentified. Now I think you have expert systems where people can identify themselves as an expert and you can use that to go and find people and that's very valuable, but you can also identify

people without them knowing it that they are an expert in something. And if they are isolated according to your organizational network analysis, you

can find ways to invite them in to be better uh connected with the rest of the organization so that they can share their information, share their knowledge better. And I think I think it's

better. And I think I think it's important to bear in mind that not all knowledge can be codified easily. Sometimes knowledge is

codified easily. Sometimes knowledge is best transferred person to person or seen in context. I'll just give you a quick example of this. Um there was a there was a series of experiments in

the 1970s with lasers in the UK. They

were trying to get a transverse uh excited atmospheric pressure carbon dioxide laser to work. It's called a tea laser. And they found that you could

laser. And they found that you could produce this laser in one lab and you could write down everything it took to

make this laser work apparently but you couldn't replicate it in another lab.

Why not? A chap did some research into this

not? A chap did some research into this doing his PhD at the time Harry Collins and he looked across what was happening with this knowledge transfer and he saw there was one particular thing that wasn't happening. you could you could do

wasn't happening. you could you could do the transfer if you'd been based in the lab and you'd seen it and then you went elsewhere then you could do the transfer. So what was happening what was

transfer. So what was happening what was happening was that the information wasn't being entirely codified. So you

know that if you have a network diagram of electronics, you know these things.

You've got lines, black lines going all over the place. You have a transistor, resistor, capacitor, whatever, right? On

it batteries right? It's like a it's like a map of

right? It's like a it's like a map of the underground in London. It shows

where things are in relation to each other, but not the distances between them. The problem was that one

them. The problem was that one particular component, a capacitor, had to be very close to the laser rather than far away. Now, it's a very heavy thing. Naturally, you'd put it far away

thing. Naturally, you'd put it far away on a bench and you'd have a a wire leading from that to the laser. But

there was a problem with having the wire being too long. That wasn't codified.

And so if you didn't see it and you didn't take that implicit knowledge with you somewhere else, you could try to recreate it from the diagrams and you couldn't do it. And this is always seen

as being a a a classic example of how very often you need to be with other people in order to see exactly how things are done in order to be able to do it properly.

So coming back to your question, I think it's absolutely possible to bring people together uh online to do things, but I also think there's a lot to be said for having

people share things in conversation face to face and see how the work is actually done. because that enables a lot that is

done. because that enables a lot that is implicit and not codified to be transferred. Yeah. Yeah. Absolutely.

transferred. Yeah. Yeah. Absolutely.

Sorry, it's a really long answer, Dominic, but I think it's really important to be quite specific that we can't necessarily solve everything electronically. Sometimes you've got to

electronically. Sometimes you've got to get together and see how it happens.

Yeah. Yeah. I love the I love the practical and the physical examples there. Um but does does this mean that

there. Um but does does this mean that when you sort of analyze and when you are involved in such enterprise implementations are there many bottlenecks or or further hurdles uh

companies face or what are even the typical issues that that that you observe when it comes to the implementation aspects that our audience could learn from? Well, okay. I mean the

the the key one is less to do with the face tof face thing and more to do with people's reluctance to be involved in sharing the knowledge they have. And

that can either be because they don't believe they have any and they they just don't they don't think they're important enough or more likely they want to hold on to it. What's called knowledge hoarding and that's true for individuals

and it's true for managers. Managers

very often don't want their teams to share what they've got outside it. And I

the when I've seen this being dealt with effectively, it's because people high up in the organization make it very clear this is part of your job and you will be ranked or you'll be performance assessed

on how well your team is sharing information. Now you need mechanisms for

information. Now you need mechanisms for doing that. It could be a proxy. Is your

doing that. It could be a proxy. Is your

team involved in this new setup we've got? Are they sharing enough? Uh can we

got? Are they sharing enough? Uh can we see them doing mini blogs or whatever?

There has to be some measure for it and it has to be given priority by people high up in the organization otherwise you will have knowledge hoarding and then you won't get those information

flows and are these you know measurable numbers like uh the number of like content pieces they create and the number of sessions they attend the number of times they interact with such

an AI. It could be it could be something

an AI. It could be it could be something like that and I think most importantly it might be something like how often because a lot of those can be gamed and you have to think very carefully about those measures. So you have to you have

those measures. So you have to you have to think about something which is as as close to being the behavior you want to have and the behavior probably is that the people are

responding to requests from outside their team for expertise. So you'd

probably want somebody you want the measure to be how often are you either responding directly or redirecting somebody's request towards a resource.

So somebody in HR wants to know more about how something is done. You either

respond to them directly or you point them towards some a resource that will help. I think that is a better way of

help. I think that is a better way of tracking it than saying have you written some stuff today? because we've seen that used quite often in the past, but unfortunately it tends to result in

people simply writing nonsense or or stuff that's very light and that's not just it doesn't have a neutral effect.

It has a negative effect because you you you increase the bloat in the system.

There's a lot of rubbish you have to get through to find what you need to find.

Yeah. Yeah. It's very interesting. Very

interesting. I I haven't thought about the psychological aspects there, but it truly makes sense. It's always the people, Dominica, or when when when it come down to the the aspect blocking and implementation, it's always the people.

I can imagine that. I talked to Catalina and she mentioned that normally the best teams to start with are actually engineers because of the nature of the job that they have to like keep sharing knowledge and educating. This is

something which I've noticed at Ericson and they've I think they're doing a really good job with their skills implementation precisely because they're very process focused and they're all interested in making the thing work. So

yeah, I think Kathleen's got a good point there. Start with engineers. Yeah,

point there. Start with engineers. Yeah,

absolutely. Um and um I also had a point around preserving conversations. What I

meant around is basically having AI accompany us to uh meetings, sessions, talks when uh it's primarily around sharing status, sharing knowledge, not

typically one-on- ones but like uh team meetings uh is that uh being widely implemented now like in our case for example we do have AI companions on on

external and some internal meetings as well. I saw it working well. But I'm not

well. I saw it working well. But I'm not sure how across the industry is is well adopted in terms of saving all those notes around even internal conversations. I think we I think we can

conversations. I think we I think we can see from our own experience of attending meetings that it is very widely done in the sense that transcripts are created and summaries are created very

widely. I don't know how much of a

widely. I don't know how much of a formal process there exists in most organizations for taking that content and turning it into stuff that's useful.

So I think that there is probably a gap there. There's something else to bear in

there. There's something else to bear in mind as well though at the moment. I think people are quite blasze about having their stuff recorded. Yeah. But I think increasingly

recorded. Yeah. But I think increasingly in the future people are going to be more aware of when a conversation has a microphone on it and when it doesn't.

And very often I'm involved now in conversations where we explicitly say at the beginning there's no recording. There's

no automatic transcript here feel free to say what you want to say because I think people are concerned that if what they say is

transcribed that can be a hostage to fortune in the in the future. Who knows

what somebody might come back and say about them in the future as a result.

Exactly. that might even uh you know reduce uh productivity or or like decision making

uh outcomes but uh can that be countered via uh sort of psychological safety in any ways by educational aspects of AI or

given the nature of the technology uh this is more about like getting used to this situation it's it's really weird um and I don't know which way it will go if you look at photographs National

Geographic magazine um had a huge issue, I think it was the 1990s, where they had a picture of the pyramids on the front cover and they moved the pyramids slightly closer together so they could fit that familiar

rectangle that exists on the front of the National Geographic magazines. And

there was an uproar about this. How

could they change a picture? Now it

might still be the case for the National Geographic, but if you look at people's photographs on LinkedIn, increasingly people are using either photographs that are slightly doed or that are completely

created by AI and nobody thinks it's odd. So we are inviting AI into our

odd. So we are inviting AI into our daily lives. It may be that we get

daily lives. It may be that we get completely used to it in the future and nobody thinks about it. On the other hand, if you look at the influence of Black Mirror, the British TV series on people's thinking, it might be people

are thinking much more cautiously about this and they have almost an at work persona and a real persona and they are constantly editing themselves before

they get their thoughts and their words captured online. Yeah. Um that's a great

captured online. Yeah. Um that's a great reference like I see the same way the the way we like naturally accept things like that would have been absolutely

weird a decade or two decades ago. Um

like what sort of things like the one you mentioned around AI editing like um the previous company I've been building was around cyber security of manipulated images and videos just getting more and

more accepted. So the it's great great

more accepted. So the it's great great example. Um and uh what's what's like

example. Um and uh what's what's like another thing where I wasn't sure but I was wondering whether you had any experience there is is keeping

communities alive and using AI there. So

basically expert communities with comm communities within an organization um is something that's that's a challenge today even like or this kind of leaves

on even if certain individuals leave the company. uh how can we utilize

company. uh how can we utilize technology potentially to uh to keep these communities thriving and enable the the knowledge sharing there. I think

the you mentioned psychological safety earlier and I think that psychological safety is is crucial in any community.

You you have to have a sense that it's okay to talk and share. Now there's no there's no technological solution to that but there are technological safeguards that need to be in place to

preserve it. So people need to feel in

preserve it. So people need to feel in the community that they can talk in successful communities that's in place as a result of the personalities who lead the community.

Um, in the physical community at Learning Technologies, one of my jobs in the opening 10 minutes on stage before I introduce the keynote is to get people talking to each other and to try in that

face-to-face environment to get people out of where they were into a space where they feel it's okay to share their ideas. That sort of thing has to happen

ideas. That sort of thing has to happen and be maintained with an online community all the time. Now if the key key people who are leading it leave then other people need to be ready

to step up for it and in particular the right sort of behavior needs to be supported in the wrong sort of behavior um suppressed and you can't have people if you have

people bad mouthing other people online or face to face it if very rapidly very rapidly has a negative effect on the whole community and somebody has to

monitor that you can do that physically or you can do it with real people. You

could do it with AI. And I think there's a lot to be said for using artificial intelligence systems, for helping monitoring communities online to keep

them not some sort of apple pie dreamland where everything is rosy, but to keep the conversation civil and to alert the monitors to to alert the

people who are looking after the community to step in and take action.

And I think if you don't do that there's every risk that the communities will go wrong. So that's I think a very

wrong. So that's I think a very practical and again psychological aspect of running community which we probably don't pay enough attention to. Um but

groups of people are not are largely selfmonitoring but they're not entirely self monitoring. They always need

self monitoring. They always need somebody I think just to help keep as I say keep the conversation civil. Once

you've got that in place, then you can have on top of it the AI solutions for putting experts together with the people who've really got a problem they need to

have solved. Yeah. Uh that's a that's an

have solved. Yeah. Uh that's a that's an excellent excellent answer like uh um I I loved how you connect the face to face elements as well. That definitely helps

against the psychological barriers we've been discussing. If we turn back a

been discussing. If we turn back a little to the resurfacing of of of knowledge and also retrieving enterprisewide knowledge using AI in

terms of implementation like what are typical like mistakes you saw like uh or or or like uh um issues that uh that

teams who've been trying to solve this encountered.

I think one of the issues is um going after the stuff that's easy by which I mean there's a there's

an old story of um Nazaret and Hodger the story in Turkey and you this story has been retold many times in in many different cultures. So Nazaret and

different cultures. So Nazaret and Hodger is found one evening he's he's underneath the street lamp and he's he's patting the floor and he's looking around very carefully. A guy walks past and says, "Nazar and Hodger, what what

what are you doing?" He said, "Well, I've lost my keys, so I'm looking for them. It's night. I want to get back

them. It's night. I want to get back into my house." And uh the guy says, "S so what? You you lost you lost your keys

so what? You you lost you lost your keys here somewhere." Said, "No, no, no. I I

here somewhere." Said, "No, no, no. I I

lost my keys uh down the road, but there's no light there. So, I'm looking here where there's light." And this tends to be the problem. We tend to look where the light is rather than where the

problem is. So we tend to try to make

problem is. So we tend to try to make the most sense of what we already have rather than going looking for the stuff that's highly valuable but it's more

difficult to find. So I think that we should be trying to capture as much around

conversation and transcribe and interpret and codify that as we can. And we tend not to do that because

can. And we tend not to do that because it's a bit difficult. Whereas there's an awful lot of information in PDFs, reports, emails, documentation, which is valuable but may not be as valuable as

the conversational stuff. And I would suggest that rather than trying to transcribe people's casual conversations, the way to get around this is to have something like focus

groups where where you have groups of experts discuss a topic that is particularly important to the organization, how they've solved it and so on. An hourong conversation with only

so on. An hourong conversation with only a handful of questions, three or four questions to prompt the conversation along. a skilled facilitator and a

along. a skilled facilitator and a skilled notetaker and also the transcribing stuff that does stuff automatically so that it's possible to really capture the expertise of those

people and codify in a way which then becomes accessible to everybody else.

It's quite rare that people do this but when they do do it it becomes a very valuable source of information. It's not

timeless though. Usually these things change over time. So you have to have a program whereby you get these people together again let's say every six months in order to just reassess how

things are going. That sort of method for explicitly looking at what people know and then surf so you're being quite clear about it. We're not eavesdropping on your conversation. We're getting you

together and we value your expertise.

We're going to surface it together. Then

that implicit information becomes explicit and then becomes codified. I

think that's the stuff that increasingly I see organizations doing. Yeah, that's

uh that's very clear. So basically uh what one thing that comes to my mind is a worry that business leaders are primary after results as soon as possible to showcase value and ROI and

now we mentioned that we really want to uh go as deep and go for the uh uncharted territories that are the most challenging. Um, how can you get the

challenging. Um, how can you get the organizational buy in for for such a thing? That's a great question. I think

thing? That's a great question. I think

you have to present people with a a very dramatic scenario. So, coming back to

dramatic scenario. So, coming back to the rail industry, right? In the rail industry, you have a bunch of people who are who have got a great deal of knowledge and who are about to leave in the next five years, they've got

retirement rates of 30 to 40% in some particularly skilled groups or they haven't got enough people to replace it.

And I think if you you don't in that case you don't have to go and convince anybody because they all know it's a problem and they're working on trying to solve it. But in the finance industry

solve it. But in the finance industry for example you have something similar um where you've got very highly skilled people but they've made plenty of money they want to retire and they want to

trans the organization should be looking at how do we get their knowledge out of their heads and into the next generation. And I think it's not it's if

generation. And I think it's not it's if you want to get the buy in, you can't present it as something that's a an academic exercise. It has to be an

academic exercise. It has to be an urgent requirement to get this done because if you don't do it, you're going to and this is the key phrase, you're going to fall behind your competitors.

And as soon as you say that, then you're going to get you're get the excitement of the people you're talking to. Right?

We're losing all these people. They're

taking our secret source with them.

we're going to fall behind our competitors. I need some budget to get

competitors. I need some budget to get this fixed. Then you're at least

this fixed. Then you're at least starting the conversation. I love that.

I love that they are taking the knowledge with them. Yeah. And they are.

They are. But that's not how it's seen because most people in organizations don't think about it that way. Yeah.

It's very insightful. Uh how is the IT security side of things from a data perspective? Like do you see much

perspective? Like do you see much barriers on that front implementing such technologies? I I have to say you

technologies? I I have to say you probably know more about this than I do, Dominic. I'm going to pose the question

Dominic. I'm going to pose the question to you. What about the security side of

to you. What about the security side of this, Dominic? Because it's a huge issue

this, Dominic? Because it's a huge issue and people are very scared of Well, I think they're rightly scared or at least cautious about it. Yeah, they are. they

are like um still it's still you know a field that's uh that's being uncovered more and more like uh I have continuous conversations with CIOS IT officials uh

and they are still banning entire AI systems in several companies like it varies across industries so the your insights are definitely so I would say

that uh for various reasons the healthcare and the pharma space is extremely innovative really deploying such AI technology is much faster than some of the other industries like for

example manufacturing. That's I would

example manufacturing. That's I would have expected it to be the other way around. I would have thought healthcare

around. I would have thought healthcare be much more cautious. I I would I would have expected as well but they they simply at least the companies that we work with or partner with have really

exerted the AI adoption by some of the more um some of the other industries such as manufacturing. I see I see a much a much slower adoption and and a lot more barriers. Um so there are

definitely differences between industries between countries as well for sure and regions who's gone who's who are who are the enthusiasts and who are the people who are more cautious in enthusiast as usual I probably the US

where you know innovation is what keeps you ahead of the competition uh but in specific EU countries like the data protection rights and and all the all

the EU policies which are being introduced and you have to comply with uh they are taking it much slower although interesting Interesting. I had

a chat with somebody last week at learning technologies who was talking about the implementation of coaching with AI and this is this is something that's going to be absolutely huge and to my surprise she was talking to two

large German companies who are going to be implementing coaching with AI at quite a lot of scale. You might not expect that because of the workers councils and in Germany. Yeah. But she

was saying it's a straightforward cost measure. You know they know they can get

measure. You know they know they can get more value for much less budget if they do this. So they're going to make it

do this. So they're going to make it happen. Coaching is really important not

happen. Coaching is really important not just for you know training people and and uh and u helping them increase their skills but also possibly for this uh

knowledge transfer. So you mentioned uh

knowledge transfer. So you mentioned uh you mentioned uh some telecommunications uh companies before. I talked to one of them in the past and they really wanted

to implement this AI coach that acts as someone's uh uh like kind of like a uh an avatar or kind of like a representative while they are out of

office. So imagine that you go for a

office. So imagine that you go for a lovely uh holiday in Spain and there is this uh virtual coach that kind of shares the knowledge that you took onto your holiday so they don't need to

disturb you on your holiday. They wanted

to implement this for 10,000 of their employees. uh they were so uh crazy

employees. uh they were so uh crazy about it and uh I think that's also an very interesting example of how knowledge transfer and uh and sharing this enterprisewide knowledge can be

utilized via coaching or via like um real time assistance in that way. You

raised something really important here.

Two things, right? Firstly, if I can have a digital doppelganger, a digital Dawn, why do we need Dawn at all? So, if

you get all of my knowledge, put it into a digital Dawn and who can answer the questions that people have from me.

Well, I can be on holiday in Spain all the time. But then the second question

the time. But then the second question is, well, there two more questions. The

second question is, what value do I add?

Well, there are some things that only Dawn can solve and digital Dawn can't solve. But then beyond that, there's

solve. But then beyond that, there's another question which is who owns the data? So, is Digital Dawn something that

data? So, is Digital Dawn something that I own as Donald Taylor or does my employer own it because they put all the work into creating and uh codifying the

data? But it's my knowledge. So, who

data? But it's my knowledge. So, who

owns what I know? Um it I haven't got an answer for this but it's something which I think anybody who works in the knowledge industries will be extremely focused on because that's where the

secret source lies. You know what's the third question there which is who has access to specific layers of the information that digital don't has right

because uh you're not not going to share the same level of uh transparency uh to uh to to a given individual as as with

your manager right it's it's different levels of information sharing so um so that's that's a very complex uh problem technologically it's a it's it's a complex problem technologically it's

also a complex problem possibly psychologically and also business-wise.

The and these are all issues which I think as we think about how we codify knowledge are going to become urgently important because this is

something that's accelerating in terms of its impact and and and the possibilities it offers. We can't get away with ignoring it. And as soon as

again coming back to urgency as soon as some organizations start to say we can do this and it gives us an edge. If you

suppose you can make your key employees available 24/7 rather than eight out of eight hours out of 24 5 days a week. That

gives you potentially a huge performance boost. So if you can do that, you're

boost. So if you can do that, you're going to do it and some companies will do it and really push the legal limits of what they can do. That means

everybody else is going to struggle to catch up and they'll be forced to try to play the game. In my turning back to the first question you asked, you would rather act as more like an FAQ bot,

right? Like uh I don't think AI is at

right? Like uh I don't think AI is at the stage where we can really replicate the way you think uh the way you act, but it's more like storing the information and getting access and

retrieving that information. So then we ask ourselves, how much of my work as a regular employee day-to-day is just being an FAQ bot. Suppose that that's what I do most of the time. Then we ask

ourselves, well, what are the best ways of managing somebody so you can get much more out of them and their creative and other potential is truly exploited and then not simply being somebody who's

answering the same question for the 20th time that week. That's a great question and also refers to a uh a conversation I had in the last few weeks which is how

we can show the path to our workforce regarding how they can uh um like learn different skills uh like apply that in

the era of AI how they can more like orchestrate these experiences rather than u you know um stick to regular

practices that that that turn them into FAQ uh machines, right? Which is not the intention. Yeah, it's it's extraordinary

intention. Yeah, it's it's extraordinary to think about that we are on the edge of something where this is going to be I think a real factor for knowledge

industry employees and the organizations that flourish will be those that do two things. Firstly,

yeah, enable their employees to get the most out of themselves beyond simply being FAQ bots, but also have the technological, legal, and psychological

frameworks in place for that to happen securely and answer that crucial question, who does the knowledge belong to. But from my perspective like uh one

to. But from my perspective like uh one thing I see continuous challenges with is this u authorization of two different layers of information. That's true for

these coaching experiences. That's true

for uh enterprise knowledge retrieval.

So, who has access to which type of information? And coming from a

information? And coming from a technology background, like that's an incredibly hard problem to solve. How do

you map that? How do you uh put in the rules? Like it's just so complex, right?

rules? Like it's just so complex, right?

Well, the crude way of doing it is to say that your access your if you think about the secret services, right? you

have access to different levels of secret information and only the very people at the top a handful of people have top secret information. You do

something similar. So you say according to your level in the organization or outside the organization you have access to certain levels of information. The

problem then is well how do you know what level of information? So you can sort out the access, but the problem is codifying the information so that it's against one of those levels of access.

And you might by mistake say, "Oh, yeah, no, that's that's something anybody could know, but it could be incredibly valuable to somebody outside the organization and it could slip out very easily." Exactly. Is this a problem you

easily." Exactly. Is this a problem you encountered as well? No, I'm just excited to think about it. I hadn't

thought about it before this conversation.

I think the points here are that uh AI doesn't just store information. It uh

reveals patterns of who uh where and what are critical and for that we can utilize all these uh um um technologies.

That's it's a really good point. What

we're talking here is about is not just the data itself but the metadata around it which enables us to understand what's important, what's not important and to reveal patterns that we might otherwise

have completely missed. And how are you seeing the adoption of these uh how are you seeing this these challenges resolved in terms of uh timelines into the future? like do you see a really

the future? like do you see a really fast adoption uh cycle around this challenge around the solutions or this is going to take a while for for the

industry to uh to really uh counter it?

I think there is this is all still so very new that I don't see anything changing the way most people do their work in the short term in established

organizations but in what are called AI native organizations typically startups anything that's been going for two or three years there's the opportunity to

take this approach and put it to work in a way that gives you an edge and so you'll the the classic innovators dilemma where established legacy organizations keep doing things more or

less in the way they've always done it so that they they can protect the revenues they've got but innovator companies come in with entirely new ways of operating based around some of the

stuff we've been talking about able to move in a much more agile way and I what I have seen is a couple of examples around this in the management consulting space so you've got huge management

consultancy companies that make very good revenues and great profits on the basis of having superb knowledge which they can access and share very very

easily. But you've got challenger

easily. But you've got challenger startups who are using the sort of information sharing pooling analyzing techniques that we've been talking about

to just do the same thing but faster and cheaper.

Faced with that dilemma, the established consultancies will almost certainly retreat up the value chain. So

they will they will sacrifice the cheaper work they used to do and allow the innovator companies to to take that and so on. It's a classic case of how change happens in organizations. Now how

long will it take to happen? I don't

know. But I would imagine the next two years we'll see management consultancy challengers coming in who are doing the same work

that your big four consultancies do but at a much lower price. And I think that will be the vanguard. That will be the

beginning of real change with real knowledge management is seen to be something that has gone from being a sort of arcane strange subject in the

1990s to being central to the success of modern organizations. So I'm saying two

modern organizations. So I'm saying two years donoring to give you a short answer to your question. I think that's a that's a you know a really pleasant timeline honestly like uh if if we can

solve uh this challenge that would be that would be that would be huge. Um one

thing that I also realized was an advice I received recently which is normally if you start with revenue generation teams and primary helping them to utilize this

knowledge that has just even faster impact on on business results and and actual revenue other than the engineers I mentioned. Uh that could be that could

I mentioned. Uh that could be that could be even like a a good tactical point uh to absolutely and I saw I saw a very simple example of this more than 10

years ago with Stanley tools I think it was in North America very simply just using short videos for the sales team to share what they were doing in selling tools from one part of the country to

another was it was distributed across the northern US and they had to bring people together to share their ideas and they stopped doing that and they just had people make and share short videos.

So it is actually information sharing but it had a tremendous effect on the sales team. What you have to have though

sales team. What you have to have though coming back to the psychological point is that the sales team has to be happy to share that information and without that nothing's going to happen. So the

technology I'm sure can do it need to make sure the sales team feels happy to do. Absolutely. Absolutely. Like I think

do. Absolutely. Absolutely. Like I think you know in summary as as we uh discussed beforehand um your organization will be successful if you can focus on surfacing uh sharing and

keeping uh the organization specific information and utilize technology specifically AI accordingly. Um I think that kind of wraps up wraps up that part

of the of the conversation. As mentioned

to you in the beginning I prepared some uh questions in addition to that. So, of

course, you are Don Taylor, so we must talk about the L &D Global sentiment survey. Um, you've been running this for

survey. Um, you've been running this for over a decade. Uh, has there ever been a result that genuinely shocked you in the recent past? Actually, yes, this year.

recent past? Actually, yes, this year.

This year, I told everybody AI has leapt last year. So, in 2024, it leapt to

last year. So, in 2024, it leapt to unprecedented heights in popularity. I

said it's definitely going down. In

2025, I was wrong. it went up even further one and a half percent more than last year. I've stopped making

last year. I've stopped making predictions as a result but that generally gen genuinely shocked me completely. It's very interesting. Yeah.

completely. It's very interesting. Yeah.

Uh I think you mentioned that as well like uh when we spoke about that a year ago was two years ago possibly. Um you

spoken at events all over the world. Uh

which country or region do you think is leading the most innovative thinking right now uh utilizing AI?

That's a really tough one because actually there's interesting stuff happening in the Middle East. Uh, of

course in the States, but as I said earlier, I think I think um the Scand Scandinavia and the Nordic countries are overlooked often as a source of innovation in this and I think there are

many reasons for it, but I think I'd say uh the my dark horse because everyone would back the US. My dark horse with AI is Sweden. I'd say that's an interesting

is Sweden. I'd say that's an interesting place to look. I studied in Copenhagen and already back in uh uh mid uh like a decade ago eventually like they've been

gathering so much data. They literally

been preparing for this. They they've

done it so in such a clever way like um AI courses specifically like deep learning kind of courses were not really existing anywhere else besides some of the Nordic universities and I think they've done a fantastic job. Why is

that? What what's what's behind it?

Because I agree but I can't work out what the common factor is. I believe

it's it's sort of uh having to having to innovate and and focus on like the the the technology industry as a whole like they all they've been pioneers for the last few decades and they acknowledge

that AI is the next thing and they saw saw the trends rising and really got the government's uh backing these initiatives like from the public

transport to other parts like already in 20 2017 like the Copenhagen like subway system was designed using several like

machine learning algorithms like that was fantastic to see eventually. I'm

doing a trip to Copenhagen at some point later on this year. I'll come to you for some tips. Yes, absolutely. You should

some tips. Yes, absolutely. You should

do that. Um what's the next question is what's something L &D people worry about way too much and something that they don't worry about nearly enough. People

worry about how things look too much.

All the research shows people will use a resource that helps them do their job better. People don't think enough about

better. People don't think enough about what's happening outside their organization. We need to be looking to

organization. We need to be looking to the wider world to see the big picture because that gives us the context to make great arguments inside the

organization. Otherwise, we just are the

organization. Otherwise, we just are the people who make nice stuff. H that's uh that's I I I I see what you mean there.

Um and as someone who is deeply connected uh within the field and the these areas of technologies, how do you stay sharp? How do you how do you stay

stay sharp? How do you how do you stay up to date? What are your what are your methods? I have a secret method which is

methods? I have a secret method which is having conversation with interesting intelligent people like you Dominic and if I have enough of those that keeps me ahead. Well uh I I I hope that uh you

ahead. Well uh I I I hope that uh you know everyone can uh can facilitate uh the right conversations and and this brings us back to the topic of organization knowledge sharing. So I

think uh connecting and networking within each company is also I absolutely think that if L & D can if L & D professionals are listening to this, they want to do one thing today to

improve their understanding of the world and the business and their personal prospects. They should go and have a cup

prospects. They should go and have a cup of coffee or tea with somebody else in the organization outside L & D. Go and

do that. Just sit down, let the conversation go, see where it goes to, and you'll be amazed what you learn.

Yeah. Um I I I see that point. Um next

one is one of my favorites, but who is someone in the learning space uh who you think kind of a role possibly that doesn't get enough credit for their work today? I'm going to say Michelle Oers.

today? I'm going to say Michelle Oers.

By the way, I have none of these questions I've seen beforehand. So this

is all off top of my head. I'm going to say Michelle Lockers based in um Australia who does a brilliant podcast.

uh it's called the um own unplugged I think it is or uncut only the uncut where she interviews real people doing real things and services great real case

studies about what's happening I mean as well as being a fantastically committed highquality individual that podcast alone is worth listening to because it just it shows the reality of what's

happening rather than the what we too often have which is the abstracted view of the world that's the reality I everyone should listen to it. Yeah. Um

thanks for thanks for mentioning that.

Um and uh maybe a final question. You

work a lot with startups and founders.

Um what's the one piece of advice you find yourself repeating to ed tech companies most often? There's one p I'm going to answer this in two ways.

There's one piece of information which every startup regardless of edtech or anything else always needs to know which is keep your eye on the cash.

Now that's a it sounds so trivial but the reason why most organizations go bust is because they run out of cash. So

I do find myself repeating that. But the other piece of information

that. But the other piece of information is that and I this is something actually you've done very well at Colossian is to focus

on the customer.

If you focus on the customer, you understand what their needs are and you see everything everything through their lens. Then you understand the problems

lens. Then you understand the problems you're trying to solve. And far too often people come to the marketplace with a solution that they're in love

with, but they haven't sold the idea to people. Rather go and find the problems

people. Rather go and find the problems that your customers are solving and help them on that way. And then everything else will look after itself. Yeah, I can acknowledge that in the first few years

as an engineer like I was just so fond of building and was focused on the technology itself but barely talked to any prospective customers and I can

definitely uh u acknowledge that and and uh I learned by doing our own mistakes but that's not to say that a passion for the product isn't important and I know

at Colossian you're really product focused but it's never enough it's necessary but not sufficient And the determining fact will always be the problems that you're solving for the

customer. Absolutely. Absolutely. Um,

customer. Absolutely. Absolutely. Um,

thank you so much, Don. It's been a really amazing conversation. As I said, it's really uh one of the most fascinating topics and and challenges that's out there today and I was so glad to basically share your learnings here

with the audience. Look, thanks Dominic.

It's always good to chat whether it's in front of a microphone in front of the public or just over a cup of coffee.

It's always good to chat. Look forward

to the next one very soon. Thank you so much and uh thanks so much for everyone who's been watching or listening. Don't

forget to subscribe to our YouTube channel or follow us on Spotify or uh give us a like on LinkedIn. And much

appreciate your um your your interest today. Uh see you on the next episode.

today. Uh see you on the next episode.

Thank you so much.

[Music]