insights! #88: From chatbots to metaverse: How you transform your business with AI

Transcript of the episode:

Joubin Rahimi:

Welcome to a new episode of insights! My name is Joubin, Joubin Rahimi, and with me today is Oliver Hoeck from digital dna. Hi Oliver.

Oliver Hoeck:

Hi Joubin. Glad to be here.

Joubin Rahimi:

Yeah, it's great. We've known each other for a while now, and I really value you and your team, because in projects you handle the things that are so incredibly important but often overlooked. Because when you're there, I know the project is set up and managed properly. Everything that happens before an implementation. Could you say a few words about yourself?

Oliver Hoeck:

Thanks for the kind words upfront, Joubin.

Joubin Rahimi:

Sure.

Oliver Hoeck:

Yes, exactly. I'm Oliver Hoeck, one of the founders and managing directors of digital dna. We've indeed been working together for a long time, we go way back. What do we do? We help companies make their digital, retail, omnichannel-oriented projects successful. And not by implementing software or selling software, but by embedding ourselves with the client teams and doing change management, project management, programme management. And these topics, making sure a project really becomes successful and also shows a return on investment.

Joubin Rahimi:

Important, right? That's the whole point. We're here today at BE.INSIDE. Great that you're here and investing the time. Now, you're also a pro when it comes to AI, and AI is a topic here today too. Were you able to take anything away where you thought, I've got new impulses here too?

Oliver Hoeck:

Yes, first of all, I find it exciting how many projects are already running that you hear little about in public. There was an interesting talk on that too. I always find it quite interesting that in the times we're living in right now, everyone talks about AI. Everyone also has an opinion up to a certain point, whether it's about the AI angle or America or whatever. There are still relatively few actual projects, even though a lot of people are engaging with it, but here I've seen quite a lot.

Joubin Rahimi:

Yes, first of all a knight's battle. That would have earned praise from you, because you also do AI projects, or much of it is about AI, but is it about AI, or what exactly is it about?

Oliver Hoeck:

Well, the discussion that's currently being held in public – the Handelsblatt has written something about it too – is that in America the question is being asked: Do all these investments actually pay off? Yes, the ones Microsoft is making, and Nvidia, and so on. It's all going towards infrastructure. In other words, do investments in infrastructure pay off financially in the end. I think that discussion misses the reality that affects us, though, because among our client base, with the companies we work with, it's firstly about the question: are there processes – repetitive, recurring, error-prone in nature – that could be optimised using AI? And the second thing is certainly topics like, I'd say, text to action, or purely text-based applications, chatbots for example. We heard that Klarna – this went through the press – is cutting 700 jobs worldwide through the use of AI in customer service, and they're decommissioning Salesforce and decommissioning Workday, because a different interface is simply being layered on top and they're now working AI-based. Those are topics that are disruptive.

Joubin Rahimi:

Are they fast, or do you think Klarna worked on it for a long time?

Oliver Hoeck:

I think they worked on it for a long time. When we talk to clients, the expectation is often: we do one project and then become professionals, and then we tackle metaverse and augmented reality topics and so on. But it's actually about, or far more important is, gaining know-how as a first step. Taking small steps first, building up know-how, and then you can take on the big topics.

Joubin Rahimi:

So the path is the goal, first and foremost.

Oliver Hoeck:

Yes, simply start. That's something we're not so good at in Germany.

Joubin Rahimi:

Start, do, try, act. Now.

Oliver Hoeck:

Exactly, now. Do it now.

Joubin Rahimi:

What are some fuck-ups you can share, that you've seen, that one should definitely make in order to really learn a lot, or also to say: okay, I won't make that one, I'll make the next ones.

Oliver Hoeck:

Well, if we talk about the topic… Let's start with AI-based chatbots. Everyone's experimenting with them. Then you often hear the complaint: we've tried this now and the results actually aren't that good. The support, the return, it's actually just not there. And this is one of the things that keeps getting forgotten: the training of the AI. In other words, what data am I feeding in, and what framework conditions am I setting for such a chatbot, such a layer, ultimately on top of the large language model, in order to achieve good results with it? And you actually have to combine that with criteria too: what's a good answer? What's a bad answer? Because that's the only way to achieve improvements.

Joubin Rahimi:

Okay. So essentially not setting expectations too high right away, and then improving step by step within the framework you've set, is that right?

Oliver Hoeck:

Yes, and another aspect too. There's a tension between a small project you start just to have a look, and yet a quality you still want to achieve. And that's why it's worth taking a methodical approach from the start, rather than diving in completely headlessly. That has a certain agility to it, of course – we just start. But if you adopt a certain sequence of methods for yourself, then you can achieve really good results even with small pilot projects, because you're able to apply a quality criterion.

Joubin Rahimi:

Okay. So then more the agile way of working, which is unfamiliar in some companies and not in others, but those who've already implemented it simply have an advantage, because they say: okay, I don't even know what the end goal is, but I take small steps and can still deviate left or right without having to think in terms of change requests.

Oliver Hoeck:

Right, right. I start, think about a path, but also how I want to get there, and adjust as needed along the way.

Joubin Rahimi:

How do you bring people along? That's also part of your task, isn't it? Yes, exactly, this human aspect always relates to two levels. One is the actual organisation in which people work, which then has to change through, let's say, software projects, IT initiatives, or entirely different topics as well. So what kind of organisation, of responsibilities, of departments, of processes do I still need access to? The second aspect is the personal level. So what does it do to the individual person, to the employee, when something changes significantly? And we serve both by looking, on the organisational and change side, at what needs to be done. And on the other side, we also cover an enablement topic: what personal skills and abilities need to be developed?

Joubin Rahimi:

I already know. We'd need to collaborate much more, also in the group context, because so far we've only worked together in the synaigy context. But I think there's still a lot of shared potential there.

Oliver Hoeck:

Absolutely. We do.

Joubin Rahimi:

Thanks for the insights. Great that you watched. If you have questions, feel free to post them. I assume Oliver, just like me, will be happy to answer quickly on LinkedIn too.

Oliver Hoeck:

We can do that. And with that … it was great to have been here. Thanks.

Joubin Rahimi:

Great that you were here. Thanks.