Skip to content
Home>All Insights>PeripherAi: The AI startup turning business data into sales

PeripherAi: The AI startup turning business data into sales

In our latest Digital Lighthouse episode, Zoe Cunningham is joined by Kim Nilsson and Ole Moeller-Nilsson, founders of PeripherAi. This startup is helping founders of small and medium-sized businesses to create a scalable, predictable and efficient sales function, by integrating the human knowledge and the data insights within. We take a sneak-peak behind the scenes at why this startup’s mission is so important and what it’s taking for this startup to grow during turbulent times.

Digital Lighthouse is a mini-series of Techtalks brings you industry insights, opinions, features and interviews impacting the tech industry. Follow us to never miss an episode on SoundCloud now: See all the Digital Lighthouse interviews online for free on SoundCloud




Hello, and welcome to the Digital Lighthouse. I’m Zoe Cunningham. On the Digital Lighthouse, we get inspiration from tech leaders to help us shine a light through turbulent times, so that we can learn, act and change as a result for the benefit of our businesses.

We believe that if you have a lighthouse, you can harness the power of the storm.

Today, I am super excited to welcome Kim Nilsson and Ole Moeller-Nilsson, who are the founders of PeripherAi. How do you pronounce that?


We were just talking about it earlier. I think we’ve converged on PeripherAi.


PeripherAi. Fantastic. So the the founders of PeripherAi. Hello to both of you and welcome to Digital Lighthouse.


Hi. Nice to be here.


Maybe could you start by telling us what PeripherAi does, and what your roles are within the business?


Yeah. Hello to the audience. My name is Kim Nilsson. I’m the CEO and one of the two co-founders of PeripherAi.


And hello for me. I’m Ole, I’m the CTO and the other co-founder of PeripherAi.


And what we’re trying to do is, we really want to help startups and small businesses do more with their sales, and help them create really scalable sales foundations on which they can grow their companies and grow their sales. And for us, that’s really about data.

Both Ole and I have worked in the industry of data and AI for the last decade and what we specifically noticed was that, especially, small businesses don’t use their data well enough to improve on and optimise their business. And since these founders and business owners told us that sales is one of their top goals this year, we felt that that’s something where data can actually help them build better sales processes. So that’s what we’re building right now.


Fantastic. And I think it’s the case that small businesses, in particular, are very poor at recognising what data they have and using it. So why is that?


I think it’s really difficult to know what to do with it in the first place. I think there’s a lot of solutions out there that, in principle, help big enterprise but few of them have actually been built to help small organisations who may not have the time, the resources, the money to hire the skills they need in-house or to get those big enterprise tools.

But I think what we’re really excited about is, how we’ve seen that in the last decade or so, SAS tools have been built, specifically targeting small businesses and being very successful by building something that is simple, intuitive and affordable. And that’s what we want to do here. Basically, take the best practice from what’s been built for enterprise and create a really scalable solution that we can offer to many small businesses.


And in fact, if you look at other industries, the SAS offerings can outcompete the bigger enterprise offerings, right? They’re accessible, but you can scale up and down.


Yes, absolutely. And of course, it’s a very cost effective way to build software, and to distribute your software. And so that’s why we can make it really, really affordable for these small businesses.


One thing that obviously is the case is that, if you’re a small business, you have a lot less data than an enterprise. Do small businesses really have enough data to be able to create valuable insights just from their data?


That’s a great question and, actually, when I reach out to these founders and business owners, they often tell me, ‘Oh, I don’t think we have enough to do anything interesting with it’. And what I really want to get across as a message here is that, actually, you need probably less than you think to get started.

And it’s important to remember that data is not just about volume. It’s not just about how many entries you have into your system. It’s also about, for example, the complexity of that data. And if we look at sales data, for example, what you typically have that’s very rich is emails or communications, that can be analysed. And that is a very rich source of data.

So I do want everyone to feel like there’s always something that can be done to get started with using your data. And you might need less than you think to get something really valuable.


To add to that. Also, from my perspective, from the tech perspective, nowadays, you can actually make more of little data and you can actually also approach it by sharing data in a better way.

And so there’s something around that, particularly using Privacy Enhancing Technologies and so on, that we are thinking of doing sometime in the future-making part of this product.

Because actually, there’s a really exciting shift happening right now in the data economy and how we use data that actually potentially helps small companies tremendously: Using data more and getting more out of data.


Yeah, super interesting. And both of those parts, I think, are each really interesting on their own.

So the first is, it’s so easy to think data is numbers or text fields that are in a database and then that is your data. I.e. unless you have data in that sense, you have nothing.

And of course, there’s a lot more than that contained within an organisation. So for example, in emails, which we don’t think twice about and we kind of send them off. It’s lost in a sense, because it’s not in a database.

So how would you go about extracting that data so that it’s usable?


Well, actually, there’s so there’s different elements to it. But SAS again, plays a big role in that, as a lot of small companies are using SAS tools at the moment. And they’re using these tools to help them with various bits and pieces, like, for example, sending out emails automatically, or, the CRM tools actually having access to the email so that they have the correspondence history, and so on.

So they already have a lot of tools that are collecting data. And I think the way to really help them is by integrating those things together. So  on PeripherAi, what we’re doing is we’re allowing companies to connect all of these different source tools together and, and basically forming a picture of the entire data set that they have, that they don’t might not even realise they have, but that they do have (even though they don’t have a database sitting somewhere, they actually have this data because it’s collected by the source tools that they’re using, and making that essentially available to PeripherAi) and then giving creative insights from that.

I think that is the kind of approach that it takes: combining things and using the different bits of pieces that you have and joining.


Yeah, okay. Emails are at least electronic. But what about all this knowledge that everyone’s got in their head?

If I think about what I do on a sales call, I’ve got learnt experience from years and years of doing it. And it’s actually very hard for me to share that with other people or pass it on or certainly formulate it in any way.


Yeah, I was going to add to what Ole said: Indeed, we often think about data as data in the classical sense of, as you say, rows of numbers or columns of numbers and letters, perhaps. And that’s what’s typically sitting in our software systems. But we should definitely not underestimate the amount of data or information that is sitting inside individual’s heads, inside people’s minds.

And especially, for example, if you’re a founder, you’ve been selling for your startup from the start, you have typically learned, for example:

  • Who is the good lead?
  • When should you approach them?
  • How do you approach them?
  • How do you respond to a certain question, etc.

And that’s indeed something that we’re also very excited about bringing into our product, through what we call, a Sales Playbook. It’s basically about taking notes of all those information that you have, all the experience you have, into the system. And the really exciting bit could be where we actually start to look for overlap between what the classical data is telling us compared to what the human’s mind is telling us.


So it’s actually much more a kind of mixed model. It’s almost integrating people with systems in a much more complex way. Yes, exactly.

Well, let’s have a chat about this other idea. You just used a phrase, Ole, that I’m not familiar with, which is privacy-. I can’t even remember exactly what it was was…


Privacy Enhancing Technologies.


It was, yes.

And I think this is really interesting, because what you’re talking about here is, as a startup, you have limited data, but perhaps there are similar startups that have different data and different insights. But actually, you don’t necessarily want to just be passing all your sales leads in this case, right? It’s very sensitive data. You just want to be like, ‘Oh, here, here you go. See if you can get some insights out of that.’ So which I guess is where privacy comes in. Is that right?


Yes, exactly. So I mean, this this Privacy Enhancing Technologies or PETs, as they’re called. It’s a new kind of technology. And there’s a lot of research going on in this field at the moment.

There are starting to be real business applications for it. And the basic idea is that you actually can do computations on data that is hidden, which means that actually you can get insights in some sense, like aggregates and averages and so on, from data without actually ever seeing it. So you can provide insights to a group of people, without them actually being able to see the data of the other participants, let’s say. in this in this sense.

So there’s a lot of different techniques – Federated Learning, Differential Privacy, and so on – that are belonging to this in this sort of umbrella term. And I think it’s really, really exciting because I think the way that, you know, we have used technology has been a bit see-sawing between a trend of decentralisation and centralization.

I mean, if you think about it, the web itself has been was a massive decentralised project at the beginning, right. But actually, with time, it’s become more and more centralised. I think there’s gonna be a new phase where actually become moved to more decentralised again, and actually, for example, cryptocurrencies is of course a decentralised movement. I think there’s always this sort of thing that happens.

And I think the way that we use data is really important. It’s really important to have have a decentralised model for doing that, because we are very concerned often about our privacy, and it does hinder collaborations between companies.

As you were saying, right, you don’t want to share your sales leads necessarily with your competitors. But it’s still useful information you can learn from each other. So I think it’s a really interesting direction that we’re moving in, I’m really excited about it, and really hope to be able to plan and build that into the product.


I think what is super exciting is this idea that that we can learn from each other and benchmark, because we often do want to support each other as small businesses, as long as we’re not direct competitors.

Imagine, for example, if we could say that companies similar to yours typically priced themselves 20% higher than you are currently doing. We want to give that sort of feedback and information to the business owner, to be able to make smarter decisions based on market data, but doing it in such a way that we absolutely guarantee the privacy of your data (no one’s ever going to find out your particular pricing or who your customers are, etc.) And that’s really exciting.


I was actually going to say exactly the same thing. I really like this idea of, I suppose it’s the kind of ‘Federation’, isn’t it, of small businesses helping each other out. So you get the power of a larger organisation and the benefits of that, while still keeping this agility of small organisations that don’t need to all do the same thing.

Let’s chat a bit about how you’re building your technology. Because you’re quite a small tech team. So, how are you structured?


Yeah, I mean, maybe it’s worthwhile pointing out that actually, you know, when we started, it was just me and Kim. We also are actually married, so it was just the two of us.

And so the tech was just me. So the very initial thing, I built myself. But then we wanted to move faster, and we wanted to, at the same time, keep a lot of flexibility. So actually, what we’ve now gone into is just hire as a number of freelancers, on the tech side, to work on different parts of the system.


I think that’s really interesting, because the perception is that you have to get your own people in, and lock them in with options, and so on, so they’re there forever and you never lose any knowledge. So what are the kinds of advantages and disadvantages of working with freelancers instead?


Yeah, I mean, I suppose advantages for us is the flexibility it gives us. So it basically means that you know, we are we are super flexible in terms of spend and what we’re doing, which direction we’re going in, and so on, adding maybe someone who has this skill set more on a very short timescale. So that’s one of the big advantages.

Challenges. It’s basically how do you, for example, create a culture, particularly on the tech side? Or how do you create a coding culture and so on. So having permanent hires allows you to shape that a lot more.

And actually, in fact, you have to. I think it’s a really important part of the role of the CTO to kind of create that culture of how you write the code best, what’s our principles behind it, and so on, and so on. And that’s a lot harder with freelancers or obviously.

So in PeripherAi, what we’re doing is actually we’re trying to make them really feel like a team, as much as possible. So that means that we actually have stand ups every day. We use Cosmos Video, which one of these online team tools, so we have a feeling that we are in the same office and so on, to make it feel a little bit okay, like we belong together.

And that actually gives at least the coherence in that sense, so that the freelancers don’t work away on their own in their own room, even just isolated on their part. But actually we have a constant interaction.


I also think the future around portfolio working, gig working, etc, I think we’re going to see a lot more of this.

Now, of course, one of the trend words is digital nomads: individuals who like they also like to have the flexibility themselves to move between jobs and locations and companies etc. to try things out.

I actually think this loosely coming together as teams, as and when you need it and you feel like it, is the future. So we’re trialling it now with with a team of freelancers. And it’s going really well so far.


Because the key really is about how you feel when you’re together, not necessarily how many minutes of how many days you will spend together. And making everyone feel like, when you’re here, you’re part of the team and we’re working together.


Well, we had a virtual snow fight snowball fight last week. They certainly weren’t part of the team then!


How does that even work on screen?


In this tool called Cosmos Video, which is like a virtual office, you go inside with like an avatar. And when you’re close to someone, video chat opens up, so if you’re walked up to someone in real life, you can start chatting.

And one of the features is there is a snowball fight. You can press a certain key and you aim at someone. Then, if they hit you, a big, big fat, snow blob comes over your screen.


That is amazing. I’m definitely going to have to look that up.

So just, finally, obviously, you’re a new business. You’re a startup.

How were you able to build a kind of artificial intelligence or machine learning business, when you don’t have the data to start with? As the the software provider, to start with, you have no data.


Yeah, it’s it’s a challenge. Of course, it’s a challenge for every AI startup. And I think one thing that always bugged me were companies who said they were AI-driven when they were not really, because you cannot have AI without having some data to feed into it.

And I’ve always been quite upfront with that. We definitely want to build our company based on AI and data and we will, but at the moment, the very first MVP we’ve built, it’s not very ‘smart’, because we can’t build those tools, those algorithms yet.

But I think the way we go at it really is to identify a problem that is really important for these business owners to solve. And something where, even if we create something that is simply automated, or simply a good way of organising your data or so already brings enough value, that they start using it.

And then over time, we start to build smarter and smarter features inside the system, based on the data they’re then sharing with us. So that with time, of course, the product just becomes more and more valuable.


We used to talk about everything as mechanical processes. So it’s like an engine or a kind of physical mechanical process. Whereas actually, now we have much more sophisticated technology, it’s much more like an organic process, right?

And actually, you currently have a business that is an infant. And so, of course, it can’t do the kind of things that a more complex organism that’s been around for a longer time, can do. And actually, what you’re saying is, it’s going to grow as you feed it.


Well, that’s how you build any product, you start with something small.

And we’ve really been trying to do this in a very agile way. So putting out an MVP just a couple of weeks ago in the market. It’s basic, it’s an MVP, it’s literally a minimum viable product. And then as we get first customers on, we take their feedback. We iterate, we iterate, we iterate.

And indeed every new version that comes out is going to be significantly more valuable than the previous one. But we feel that’s the best way to build a product because you’re actually building with your customers essentially. And not just something that you think for some reason should be a great product. So yeah, we’re encouraged with the progress so far.


Yeah, you’re getting that data from what your clients need and what your clients want, and you’re using that to shape what you build. That’s fantastic.

Thank you so much, Kim and Ole for coming on and helping us to shine a light for others.

Kim & Ole: 

Thank you for having us.

Data Engineering

Streamline Your Data Journey: Discover how our expertise turns complexity into clarity.