EP126 – Becoming the Chief AI Officer for Your Clients with Nils Howland

Hello, and welcome to It's a Numbers Game.

Today we are joined by Niels Halland, co-founder of Ironbridge, an AI first fine-build focused

on partnering with MSPs.

Niels brings a rare combination of perspective to this conversation.

He's a trained computer scientist, has a background in M&A at JP Morgan, has scaled businesses

at serious speed, including one from seven to seven hundred people, and has also run

an IT services business himself.

More recently, he's been investing heavily in AI and security businesses across Europe,

deploying around 300 million euros into high-growth technology companies.

What makes this conversation particularly interesting is that Niels believes MSPs are uniquely

positioned to become the delivery platform for AI in the SAP market, and that doing so

requires a fundamental rethink of how MSPs operate and how they charge.

So today we're talking about why AI is different from previous tech ships, what customer

zero really means, and whether MSPs need to become something entirely new.

Niels, welcome to the episode.

I don't know.

Lovely to meet you.

Happy 2026.

Happy 2026.

Yes, we try not to date these recordings, but yes, we should be okay to release this during

2026, and therefore we'll be okay.

So Niels, you and I met last year and had some really interesting conversations that we definitely

wanted to share with our listeners on the podcast here, and your background down is an investor

and an operator of businesses with your vision of what the AI revolution will mean for the

MSP community, I think has some real inspiration.

In terms of kicking off the conversation, perhaps you could tell us a little bit about where

you see the MSP development happening in terms of AI, first of all.

Yeah, I mean, before that, I would quickly start with an introduction on myself.

One of the co-founders here at Ironbridge, we are in AI first, by and build in the UK,

with a global ambition currently focusing on the UK market.

Quickly on my background, I'm a train computer scientist with a background in M&A at JP Morgan.

I then went on to operate a few businesses.

One was a fast-scaling journey, so hyper growth, scared of business from 7 to 700 people

within 12 to 18 months, and then I run our family business and my two services in Germany,

which is where you might hear my accent coming from.

So we scared that business from single to double digit growth.

Yeah, then I had to great privilege to work with a phenomenal set of founders over the

past couple of years, deployed around 300 million into Europe's most ambitious startups.

And as part of that, we're Greg Chosky with both AI operators, as well as IT service operators.

So Dawn Capital, the fund that I've been working with together, has been the early investent

to MIMCUS, which many of you may be very familiar with from an EMS security site, but also

over time, some of the investments that I've been leading have been LeanAX, which was

an IT enterprise architecture management tool, which we exited to SAP, BODGARD slash Black

Wall, which is a network security provider that works with many MSPs and hosting service

providers as well, as well as a bunch of other providers that have worked on building

the autonomous AI first security operations center to enforcer, which was one of our latest

deals that I've worked with very closely with the team.

So dead a bit about my background, quite frankly, what we're up to now is we've realized

that there's a massive opportunity outside of the pure play software bid.

And we will leave that defensibility and differentiation in the long term will come

from the trust and relationship with end customers.

And MSPs have an absolutely phenomenal insertion point of AI for the end customers.

We have a natural partner for the end customers to deploy AI over the coming years.

And we believe that's a great opportunity.

So we want to work with MSPs, want a bunch of different things, firstly, essentially,

changing the operating model internally, allowing the MSP to scan much faster without adding

additional headcount and being an AI first organization.

So thinking about it as being your customer zero.

And then the second thing is really helping the MSP and working with the MSP together

on selling AI outcomes, automations, end-to-end, which requires a change of the commercial

model, quite frankly, but you're all of that by investing into MSPs and then working

closely with them.

So we consider us investors and operators at the same time.

But that is as part of a quick introduction and super happy to jump in and why we're so

excited about MSPs outside the fact that we've been working with the industry over the

past couple of years already.

That maybe let's start with the end of mine then and talk about the commercial model

for MSPs changing and then work back from that.

Presumably, that's the opportunity that you've seen in terms of new revenues and working

relationships.

Yes, absolutely.

So briefly, just to set the scene, I guess everyone would be aware that it's a unique

point in time for MSPs and AI forever changes how we get work done.

It's going to be the most important industrial revolution yet and it is fast in terms of

adoption and it has a greater impact than previous industrial revolutions.

What we've learned is that adoption is painful and slow.

Change is the thing that we will have to work on quite frankly.

And what it will require us to rethink operating systems from fast principle.

Which means we can't just automate existing work but what we should really think about

is how can we actually unlock net new work.

So provide services to our customers that haven't been possible before.

Quite frankly, that now possible with AI.

Free up time within the team to spend more time with the customers instead of spending

time on filling out tickets and or different items in your CRM.

Yeah, I think most importantly, we believe that MSPs are the platform that will deliver

that revolution and that change to the broader SME and SME space.

And there's no more natural partner.

They are the McKinsey's of the world.

They are the large system integrators of the world.

But they don't really understand the SME and the SMB as well as MSPs do.

They don't have the trust.

They don't have the context in order to have the relationship.

And again, MSPs do have, they own their tea state quite frankly.

They have access to the technology to the infrastructure and to the data.

Which means that they are the best sort of partner on their change.

Same as they've done over the last two or three decades,

handholding their partners through the technology,

revolutions and transformations.

Which is why we think, yeah, it's a phenomenal time for MSPs not just from an internal

perspective and transforming the internal model,

but more importantly, starting to sell AI solutions to end customers.

But to dive a bit into your question, what does it mean?

It means that you will need to change the tenants of your commercial model.

You will need to move from attacker to a strategic partnership with your customers.

From selling times and materials to selling outcomes,

which is very much a term that is sort of being thrown around the industry.

But how you start with this is sometimes asking a customer,

how much are you spending on doing that work today?

And how much would you be willing for willing to spend for that work to be done automatically?

Like, what's your, what are you willing to pay quite frankly?

And then with that, it's an attribution problem that you will solve over 10%.

But yeah, in the end of the day, all of that means that you will need to start much earlier,

sort of with a problem discovery and touching on business problems quite frankly and

business outcomes with your end customer, which will allow you to go, yeah,

11 beyond IT infrastructure into the boardroom almost.

So the real goal that we have is to become the chief AI officer of SMBs and SMEs

and how Jensen from Nvidia framed it early February last year was you become the HR

of the AI agents within your end customers.

That means eventually you will be the person that sets up on boards that manages,

that orchestrates, that monitors, that evaluates the jobs that the agents are doing around

your business. And as a CEO, you will need a partner that tells you what on earth are those

10, 15 agents doing that are running around my business? What data are they touching?

What is the RI of those agents? Are they still performing? Have they been recently updated?

Are they running on the latest model? All those questions will be relevant and important

to be answered as the strategic AI service partner. And we believe it's the next evolution of an MSP.

Be honest, when was the last time you actually felt in control of your MSP finances?

The late nights chasing invoices, the cash flow stress, constant pressure to grow profit,

it can feel never ending. At MSP Finance team, we've built the complete outsourced

finance solution just for MSPs. From day-to-day bookkeeping to annual tax, management reporting,

and full virtual finance team support, everything you need to run a profitable,

stress-free business is under one roof. So whether you want to regain control,

free up personal time, or finally build toward a real financial freedom,

we can help you get there faster. Visit MSPfinanceTeam.com and discover how our outsourced

finance team can help you stop surviving and start thriving.

So to pick out two things that I heard there, to help me understand the first thing I heard was

that the primary commercial model, the pain that the end client has, is one of

saving time and therefore reducing staff costs. And then the second part I heard was that

although MSPs today have been focused around, as you quite rightly said, predominately infrastructure

management, as only some then venturing to the line of business application and beyond

office productivity, you're then saying there's a new infrastructure requirement that's going to be

created around AI on top of the traditional IT infrastructure. Effectively, we create some new

problems by solving one before we have an agent to do things for us, but then the next problem is

managing those agents. So yeah, maybe pick those two points if you would.

Yeah, I think the first point, it's not just around reducing costs, it's more around really

making sure that your core team can spend time where it's most valuable, which means spending time

with your end customer instead of spending time completing redundant mundane tasks on the

daily basis. So we want to shift the focus to a coming higher quality, to becoming faster,

and in a sense, cheaper delivering the servers,

whilst you actually have the opportunity to spend more time with the customer.

How it looks like from a perspective of what we want to build is we want to grow much faster

at the same hard count. And there's some players in the market that are already showing that this

is possible. I believe some of the examples you've covered in previous and previous sections

of this podcast where people have talked about the natural insertion point of AI into your help

desk, where it's around triaging off tickets, where it's around some auto responses on level one,

where eventually you will need to make sure that only the tickets that are high complexity and high

value will actually reach a human being and are those where you know that the customer relationship

is critical at this point on time that you will need to inject. Now, there's lots of other interesting

examples and we could go through the whole organization, lots of them will be, we'll also be back

off as tasks. But many of the MSPs are already on the journey to being as and or are already a

strategic partner, which means that lots of times spend on creating reports, creating quarterly

business reviews, meetings with the customer to play back or really learn about the business.

There is a function of gathering data across a disparate set of systems

from your cross-strike or huntress to your sort of take a management system to your

learning systems and those tasks currently take a lot of time to be prepared and take humans then

to be performed. So, ultimately, what we want to achieve as well is reducing time in these

preparations and really making sure that it's more on the personal interchange rather than the data

cleaning and getting everything from different points. Those are just two examples. There's many

other examples when you think about professionalizing your sales motion and go to market and using AI

to be more efficient on that perspective and to have high number of leads in and have essentially

high velocity in your sales motion to move to a high velocity motion in itself as well as finance.

But, long story short, lots of things to be done on the first bit. On the second bit, the important

thing again is to change and it will be a process that will happen over the next year to come

to change the way you sell and to change and focus more and more onto the strategic topics

and the business topics and how you typically start the conversation is understanding what the

bottlenecks are in your end-customers business. Ultimately, tying those to the P&L of your end-customer,

which is, hey, your end-customer, the CEO really cares about one thing is top-line,

just the revenue that comes in and how can I get more of that? And then it's the bottom line.

How much could I do? I was just going to ask, can you sort of already sort of started to surface

this subject? If we go back a couple of decades, MSPs have had to

check and pivot as technology has evolved. There were networks, first of all,

then cloud, then security and AI, if we're summarising it at a low level. And each time the MSP,

to some extent, has had to kind of review what they are. Are they a box shifter now? Are they now

a service provider? Is someone else a service provider? And I'm just an interface. Oh, hang on,

the security thing. Now, my security first, oh, hang on, am I AI first? What's going on? So,

how do you think the end user market perceived this latest change?

And are they thinking, hang on, AI is something different to generic infrastructure and security

management? I need someone with a different set of skills here. They can't just pivot from

supporting 365 one day to suddenly being business process experts and understanding my business.

So, I'm going to choose someone else who's not got that legacy MSP background. I'm going to choose

some kind of developer, some kind of business consultant type credential. So, how do MSPs

need to kind of, I'm not saying this is going to be the case. I'm just suggesting it might be

the case. How do you think MSP should approach this? Should they look to redesign what they are

fundamentally or make some kind of gradual shift from infrastructure security into AI, AI,

HR manager, as you said, then into business consultants? I think I should put this question back

to the audience. How many of the MSPs and IT service providers in the audience have been approached

by the end customers, asking them what they can do with AI? Whether it's a simple question around

co-pilot agents or whether it goes beyond into, hey, can you help us set up some AI systems for our,

go to market for our CAM for our Microsoft Dynamics or whatnot. We've done a survey earlier this

year and I don't know the results on top of our head. It's on the dawn website where we have

asked MSPs and the majority said that customers are approaching them for advice with regards to AI.

So, to answer the first set of the questions, I believe that end customers are looking to the MSP

because the MSPs are the natural point that have guided them through technology change over

the last decades and the cloud was the same move essentially. It was a completely new environment

for end customers. It was change management that was required. It was painful, but we see

particularly in markets like the UK and the US, it's extremely high penetration, markets like Germany

where I'm from a bit slower, quite frankly, on their shoulders. Now, in the second bit in terms of

what these should and should not do, in the end of the day, it's the decision of every single MSP

as to how they want to embrace this change. And we would have seen the same, again, each of the

times in the history where we've seen movements from people changing their pricing models,

from people sort of moving from box shifted to service providers, people evolving from a people

like service, managed service provider to manage security service providers. Those are active

decisions that you as a business leader have to take for yourself and where you have to decide

and where you need the future. We want to partner with people to be super upfront and transparent

that are extremely excited about this opportunity. And this opportunity means, again, change

on all fronts, which means that we need alignment with the entrepreneurs that the operating model will

change. We will spend more time with customers. It would be more of a consultative sales fashion.

All of that on steroids with regards to AI and using as much technology as possible to cut

the time during knowledge work. But most importantly, we will increase the time spend with end

customers. And that is sort of high level. The focus that should be in the prime light.

It does seem to me, though, that, and I agree with what you said, but it seems to me that there's

something a bit different this time around around the business consultancy skills required. So

until now, a tech first tech, you know, an owner who was an engineer as a tech, that's how he built

his business. He's in his comfort zone around everything that they do with their clients.

It's all been around tech, ultimately. And now we're asking them to be business experts as well,

which we've always, to some extent, wanted them to be right to have them having the right

conversations and setting outcomes still, rather than tech. But is this something a bit more

fundamental going on here that's going to be an opportunity for some and risk for others?

Hey, I think it's the right question to ask. And again, you would see that some have moved

from being a pure play technology provider into, actually, I'm your virtual size or CIO, whatever

you may call it. I'm in more time with you and leadership and or I drive the roadmap together

with you. But all of that tied to what does it actually mean from an RI perspective? So that was

the first change on that side. But hey, look, that's why I think it is quite a unique opportunity

for us to partner with MSPs because we as Ironbridge bring exactly that to the table.

We will bring an autonomous, independent unit, a team that runs on a separate P&L that will

support you both with a perspective of how do I change my internal operations and

did something that skates much faster, essentially. And then with regards to how do we commercialize

and how do we sell outcomes? The model that has has been proven in the past and that has worked

extremely well is the forward deployed engineering model of Palantir, where they've shown that

it's engineers and technicians that are deploying into their end customers and solving business

problems. And what they are just doing is they're extremely curious and they dive deep into the

problems that their customers have. And then they solve it with technology. So that is a beautiful

model to look at. The Palantir model, just a wish level. Sorry. It's from Palantir, which is,

I suppose the stock has dropped a little bit, but it's one of the richest multiples on the

public markets. They are in the end of the day, they're a service business, which is software enabled.

But yeah, that's exactly why we want to partner with MSPs and support on that front.

We have collected a range of operators on our cap table ranging from people at organ AI,

to people at hugging face, to people at deep mind that are supporting us to one being at the table

when new things are coming out, but equally working with us through the challenges

that come with changing the way that we sell. Because what we see from software side is that

software providers realize they can't just purely bring software to an end customer anymore,

but they will have to also attach the services around it. And they realize that they will

need to go much deeper into the engagement with the customer than before, because they need a

different set of data. And it's not just out of the box software anymore, but eventually it will

be sort of a custom software set for your end customers, which again, that's sort of the

straddle from both sides, right? I see it from a software side and from a software side.

Yeah, pretty interesting. And I think one of the, one of the things that always stuck in my mind

and back to the start of my career was conversations around how much is invested in software versus

the services. And I'm pretty sure it was like best practice was it was like you spent as much

on service and implementation and support and optimization as you did on the product itself.

And it in effect, if you're a software provider, you've been leaving half your revenue on the

table as a result of that. Unless, of course, you've got a really engaged partner community that

are your implementation partners, which of course is sort of the MSP model really taking a

much of product, for example, and then providing a service solution wrap around it and then taking

that to market. And just to Adam's point around the end users view of the MSP and the authority

that the MSP has to position themselves as now a business process, optimization, consultant,

really is the fact that they'll have done it themselves on their own business. I think is Nils' point.

Yeah, absolutely, which is why we believe in being the customer zero, starting with yourself,

understanding your processes, workflows and business and tying it to the business goals. And

then you go to your customers and say, Hey, we're going to work through processes through your

workflows and tie them to your business goals and be a strategic partner. But it's a journey.

And we've seen many MSPs already on their journey over the past years. I think the most interesting

movement in his sense is MSPs moving to being an MSSP, but now also thinking around how can you

on the Microsoft side deliver dynamics, which nothing else from a perspective of how you sell it,

right? And Microsoft has shown with the products that they're releasing, they're sort of giving us

the journey that they want their partners to go in as well, which is Microsoft has focused a lot on

on power automate. Now they've pushed dynamics a lot and now they're pushing co-pilot, right?

All of those essentially demand you as a service provider to spend more time with your customer

and their business rather than to stay in the background and being a strategic partner,

which is great because ultimately that means that you have a unique selling proposition and there's

a time in the market where the market will reshift and you can actually differentiate and say, Hey,

I have a service that I deliver that is differentiated compared to all of my competitors that are

just commoditized offering and or offering perhaps better SLAs. That's how we that's how we

think about the world. But yeah, we are right. The unquoted number that we've heard is always that

every dollar you spend on Microsoft, there's around six to eight. Some people say even over 10.

The number that we keep getting six to eight attached in service is spent. So it's a massive

opportunity to mess with my goodness. Yeah, right. Okay. And so even more more opportunity than I

thought. And what one of the one of the other things that strikes me about the last phases of

of technology like if you take the in isolation, the line of business application, which which

of course could be dynamics could be power automate. One of the key challenges around that has

actually been the constraints around good quality consultants to actually deliver that. Oh,

and of course, AI is the answer to that because now you can create agents you don't have to rely

fully on people to provide that consulting delivery services. It were yeah, 100% and look a lot of

the time spent is not just understanding the business and mapping business processes, but it's

essentially once you've gathered all of the data, it's then digesting the data and both the

data gathering and the digesting of the data AI will be massively helpful. We've seen it ourselves,

we've worked with some sort of mid market customers to to prove the AI journey and we've

built lovable interfaces quickly and just the data we've prompted it with AI and got it sort of

all contextualized, which just allowed us to do a number of interviews within and across the

business in a time that was much shorter than if we had to be there in person, run all of the

interviews right under transcripts, then map what we've learned in each year, sort of use our pen

and underline what was the important things they've set, know which all we'd run that all through

through our AI in the end of the day to get that all maps into what are the bottlenecks of the

business and from those bottlenecks, what are the agent workflows that you would want to build

and suggest and then the important bit though for us was to say, hey, now we are going to prioritize

because we understand the priorities of the board and of the leadership and those two things,

bridging those two things, neither the context that we had from the previous interactions with

leadership and the board of these businesses. And they agreed with only you saying, just on that.

So just some numbers, roughly how many people did you interview then through AI and roughly how

much time did you save on that particular exercise? Yeah, I mean in that specific case it was around

15 interviews that we run. We run, it's a good question how much time it would have saved because

we didn't actually do many of this. But I would assume that this would have probably taken us two

days of analysis and we were done within I think it was two hours. It was super, super quick.

The most important thing in those two hours was then once we had all of the data at our fingertips

in a summarized and consolidated categorized version was to say, hey, how do we prioritize it?

And then exercise, we sort of did ourselves prior to then engaging with the leadership,

but when we had to see level leadership meeting, we came with a suggestion and we just discussed

what is important to them and what isn't. But we hadn't really had to discuss all of the

in the 1500 data points in what but we could jump right into it, which was extremely helpful.

And thinking back to my earlier question around the ROI calculation,

you've got time saved, you've got the speed of which you can arrive at the next phase of the

process. And you've also got the fact that you can be doing higher value tasks for more of the

time than the lower value tasks. So I guess in my mind, the ROI calculation could be multi-faceted,

you know, we're saving, I don't know, 10 hours that would be, that would have a cost of 50 pounds,

but we're also now enabling those 10 hours to be 300 pound an hour task rather than 50 pound an

hour task, for example, is that the sort of process that you might go through? Yeah, I mean,

a hybrid with outcomes based pricing and thinking about ROI is the attribution problem at the

moment, which is going to become better and better over time, the more you spend with your

customers in production, which would then allow you to actually attribute it to the essential

bit more with me by attribution problem. Attribution problem means how much time did it in the end

actually save? Okay. And making sure that this data essentially matches up, which is why

in the first place, what we're again, what we're saying from respect to this outcome-based

pricing is the fastest takeoff is bottoms up. You really need to understand the most painful

problems and they tell you how much it's worth to them if you ask them. We literally, in those

15 interviews, when we gathered all of the use cases, we started with bottlenecks and then

gathered use cases, right? Out of those bottlenecks. It's like, what are the bottlenecks in your

business in your line? Like, what keeps you away from being faster or better or or deliver better

service? And then we just simply asked them, how much does this worth to you? And that may sound

very unsophisticated now, but we came up with a number at the end of the day. And it was quite

big and the sea level was extremely happy about it, because that's what the team was willing

to spend for a century solving these problems. And then all of a sudden, it's a very different

conversation. Over time, you will have to, you will have to move more and more into data driven

attribution as to what exacts part of the business is touching and how much impact does it actually

have on your top line, on your bottom line, which is a function of exactly what to say down,

time saved and customer feedback. Brilliant. So I think my takeaway from this, don't overcomplicate it,

ask some very simple questions, direct questions, you'll get simple answers and the client is likely

to respond and say it costs us this. Therefore, that's your budget. Can you fix it for them within

that budget? And away you go. Yeah, it's just get out there and start doing and you will learn so much

with your customers and your customers have trusted you for the last 10 to 20 years, which is the

beauty of this industry, right? And why is that not just because MSPs have been a reliable

partner, but also because it's a great bunch of people. And they probably like you as a person,

and they probably like to experience a bit of sort of a new what can we do and what can we change

and are excited about it. So start doing, caveat that it's early days in some aspects, but people are

happy for you to be practical is my big learning. And then it's all about like spend more time with

your customer rather than theorizing to spend time with them on the ground and you will learn from

them what the biggest challenges are. And again, but they also really to pay to solve to resolve

those big bottlenecks in the business. Yeah, I really like your point there about the caveat,

which is it helps set expectations and take some of the pressure off so that you can be

perhaps more more inventive and open to change rather than playing it safe and not taking a risk.

So yeah, no, as as always talking to you now, it's a fascinating conversation, very inspiring.

It's around this time in the episode that we offer our guest a shameless plug. So if anyone wants

to carry on the conversation, how best to get in touch? Yeah, that's a great question. I mean,

I'm on LinkedIn. It's the best way is just to connect me there and or to pop me over in

England, like email is near to ironbridgesp.com. My co-founder Owen will be coming onto this show soon

as well. He's very opinionated with regards to some of the questions we've tasked upon ROI

commercialization model and he's sort of our AI mind. So there would be lots of interesting

conversations to be had with him as well. And yeah, get in touch. We love to speak to MSPs and

hear what they've experienced and share some more of what we're learning and experiencing. I've

I've probably met 200 plus AI businesses in the last two years. So I've met most of the AI founders

in Europe and lots of interesting war stories from them from the software side and met many MSPs.

So so always keen to to exchange and swap notes and share some of the learnings on both sides.

Very good. There was it's been a pleasure talking to you today and thank you very much for

joining us. Thank you so much.