27 August 2026

From Copilot to smart apps: AI within IT Local





How do we help our clients using AI? And how do we apply AI ourselves? This article provides an overview.

AI… everyone’s talking about it, and it’s a hot topic in the SME sector too. As an entrepreneur, you’re bombarded on LinkedIn with the latest AI innovations, success stories and critical voices. And perhaps you’re already experimenting with AI yourself, or your staff are using various tools. This quickly raises questions. Which new possibilities are truly relevant to your organisation? How do you get the most out of them? And how do you use AI in a responsible and secure way? We’ve noticed that our clients are also exploring this area: on several occasions, we’ve been asked what exactly we, as an IT partner, are doing in the field of AI and what we can offer them. The answer to that is quite extensive. So this is a good opportunity to provide a clear overview of our AI applications, services and our own experiences.

AI is everywhere. This was also evident following a round of discussions at IT Local, where we gave Gaby, Bart, Matthijs, Paul, Jenna, Robert, Terrence, Sunny, Pierre and Chris the opportunity to share their expert views on the subject.

Microsoft 365 Copilot is probably the best-known example of an application, but AI is now featured in many more areas of our work. It helps to secure our clients’ IT environments, enables the development of smarter apps and supports our own staff in their day-to-day work. At the same time, we do not want to make AI any bigger or more complicated than necessary. Not every form of automation is automatically AI, and not every process is necessarily improved by adding AI to it. That is why we focus primarily on the practical value: what problem does it solve, what benefits does it deliver, and what remains the work of humans?

In this article, we outline which AI applications we currently offer, how we help clients get started with them, and how we use AI within our own organisation.

In this article, you’ll read about:

1. Copilot: the most visible form of AI
2. AI in our existing IT services
3. Smarter apps and processes with Clevir Solutions
4. How we use AI ourselves

1. Copilot: the most visible form of AI

Microsoft 365 Copilot: AI within the familiar working environment

For many clients, Microsoft 365 Copilot is their first real introduction to generative AI in the workplace. Copilot is included as standard in the Office licence, in a limited chat version. With a paid, more comprehensive version, it also works directly within programmes that employees already use on a daily basis, such as Outlook, Word, Teams, PowerPoint and Excel.

We supply the licences through IT Local. In addition, we help our clients determine whether Copilot is right for them, how they can get started safely, and how staff can actually learn to use it. Customer Success Team Leader Gaby Phielix is happy to tell you more about this. She begins by explaining from her perspective exactly how the Copilot licences work.

What licence do you need to be able to work with Copilot?

Gaby: “If you have an Office licence, you get Copilot Chat included. It’s similar to ChatGPT or Claude; there’s no licensing structure involved. That is, however, linked to Copilot Premium, the paid basic licence, which allows you to use Copilot within your Microsoft Office environment; it appears in all your applications and in the M365 chat, and you’ll also be shown whether you’re staying within your own work environment or going to the web. Within your work environment, via Microsoft Graph – as this ‘engine’ is known – you have access to all the data linked to your Microsoft account and to which you’re entitled… email, appointments, Teams chat, and OneDrive and SharePoint documents. And it’s secure, because you remain within your organisation’s cloud environment. That’s handy if, for example, you want to create a summary of a document. In the free version, you’d have to copy the document, paste it into the chat and ensure it complies with the GDPR. With the Premium version, you don’t need to do that; you can simply ask: ‘Summarise document X’, and it will do so without me having to worry that my data is being used to train the model.”

Extensions for specialists

“In addition, there are extensions to your Premium licence, such as Copilot Studio, Copilot GitHub and Security Copilot, which are intended for developers or security specialists, for example. Copilot Studio is also of interest to people who would like to build their own company-wide agents. For these extensions, you can purchase additional licences or use them on a pay-as-you-go basis, in which case you buy a sort of token that you can use to access them.”

“Regulating AI from within an organisation would be better than just letting everyone do as they please”

More use of AI…

More and more organisations are taking an interest in AI. Gaby: “Since Copilot became available to medium-sized SMEs, about three years ago, a few clients have already signed up. But in many organisations, I’m now seeing that employees have started using AI on their own initiative. That’s not very secure. It would be better to regulate AI at an organisational level rather than just letting everyone do as they please. What I’m also seeing around me is that AI is being integrated into many existing apps.”
From Copilot to smart apps: AI within IT Local

… and a focus on AI literacy

Greater use of AI also means that attention needs to be paid to so-called AI literacy. In other words: how it works, how to deploy it and how to use it effectively and safely. Customer success can help with that. Gaby: “Absolutely! Since the launch of Copilot, we’ve been keeping a close eye on it. Initially, the focus was on how to use it – that’s what our first webinar was about. We then looked at how to train people in its use. During our user training sessions, we cover the practical applications of AI and answer questions such as: how exactly does AI work, where does AI get its information from, how do you use AI safely, are roles and permissions set up correctly, and how do you write a good prompt? But we also cover where you can find Copilot in your Microsoft environment, how it works within the apps, and how they work together. To clarify that last point: from within your Word document, for example, you can set Copilot to work to create the basis for a matching PowerPoint slide deck.”

A suitable Copilot training course for every situation

“In our training sessions, we tailor the content to the specific Copilot variant the client is using. The most enjoyable sessions are those for Premium users, where participants can get actively involved straight away, but we can also provide training for the free version, in which case we focus more on prompting and the safe use of Copilot.”

We also provide training to companies that are still assessing whether a paid Copilot licence offers added value. After all, it is quite a pricey licence, costing around €30 per month, and you have to pay for a full year in advance. Incidentally, there are also more affordable options for educational or non-profit organisations. Furthermore, Microsoft offers a discount on a Copilot licence if you also purchase Business Premium."

Allaying fears about AI

The benefits of AI are certainly recognised, but there is also scepticism amongst some users. They are also aware of reports concerning excessive electricity and water consumption, a dumbing down of knowledge, copyright infringement, privacy breaches and uncontrollable situations. We won’t be addressing all of these in this article, but we will look at the immediate concerns surrounding the use of AI and how we at IT Local can allay them. Gaby: ‘The biggest fear companies have is that something will happen to their data. That’s why we need to start from a secure foundation. We help with this through our Copilot readiness scan (see further on in this article, ed.).’

“Copilot is there to support you, so you don’t have to start from scratch”

“I’d love to see everyone using AI because it makes your work smarter. One of the most frequently asked questions I get during my training sessions is: will AI take over my job? I say no, but your work will change. For example, moving from doing everything yourself to focusing more on evaluation. Copilot is there to support you, not to make you start from scratch, or to help you improve your work. For instance, I always have my texts checked for spelling and grammar errors, and when I’m preparing a presentation, I ask myself whether I’ve missed any topics.”

Working smarter with AI agents

“Another thing we do is, once users have learnt how to use Copilot, work with them to see how they can work more efficiently with a personal agent – and by that I mean without using Copilot Studio. Think, for example, of an agent that can organise my to-dos for today into a handy list from all the Microsoft apps I use, such as Outlook, Teams, Planner, OneNote, To Do and Loop. Or awriting style editor agent that I’ve created for myself, for example. It rewrites my texts in the IT Local house style. We’re also looking at how we could deploy agents organisation-wide to support work processes. We’re doing this in collaboration with our engineers.”

Want to know more about Copilot agents?

Take a look at our two webinars on this topic:
▶️ Getting started with Copilot Agents
▶️ Copilot Agents for SMEs

Safety first

Gaby has already mentioned the importance of the safe use of AI on several occasions. Engineering Team Leader Bart Hunik cannot emphasise this point enough to clients either. It is no coincidence that he is the initiator of the Copilot readiness scan offered by IT Local.

Bart: “If I were to make a rough estimate, I’d say that around 10 per cent of our clients currently have a paid version of Copilot. I find that rather disappointing, given the risk you run when you allow your staff to use the standard version or other free services. Because as soon as staff start entering company data into those services, it becomes available to those AI agents. And then, in effect, that data leaks out. The advantage of the paid licence is that the data doesn’t leave the tenant.”

“Data can leak out, but it can also leak within your organisation”

“Incidentally, that security works both ways: so you have data that can leak out of your organisation through the use of public agents, but it can also go the other way, meaning that certain data can leak into your organisation. If you have documents drafted using free AI tools, it is possible that unverified data sources may be used and end up in those documents. For example, if you want to create a guide on how to set up MFA and you have Copilot write it, Copilot might draw on a source whose content is incorrect, meaning your guide will also be inaccurate. If this happens frequently, the quality of your internal documentation will decline.”

The importance of the human touch

That last example also makes it clear why human oversight is so important. Bart: “It’s always necessary. I compare an AI agent like that to an intern; it doesn’t know everything, so the ultimate responsibility lies with you. And you must always check the output against your own expertise. I think the human element is more important than many people realise.”

This human aspect also receives a great deal of attention in the Copilot readiness scan offered by IT Local, which Bart initiated.

Bart-copilot-readiness-scan

Copilot readiness scan

Our Copilot readiness scan shows where you currently stand and what steps are needed to deploy Microsoft Copilot safely and effectively within your organisation. A technical scan highlights your risks and areas for improvement, for example regarding access rights, information shared too widely, corrupted data and information management. Through a supplementary analysis based on an interview, using the ADKAR model, we examine the human and organisational aspects: to what extent is your organisation ready to start working with Copilot, and what level of adoption might be necessary to ensure its successful use?

What was the reason for developing the scan?

Bart: “Mainly to check whether AI is safe to use; that’s actually the technical bit. By running a scan on your organisation’s tenant, we check whether it’s secured or freely accessible, and whether you can access and use whatever data you like. With the scan, we check, for example, whether the permissions structure is correct. Because AI utilises the existing permissions structure within your organisation. If your staff have too many permissions, they’ll have access to sensitive data. Sometimes you’ve set certain permissions too broadly, which only become a problem when AI is used. For example, if you’ve shared a salary folder with the whole organisation and nobody knows the link to that folder, there hasn’t been much of a problem up to now. But if you now use AI to ask how your salary compares to that of your colleagues, things go wrong and you can immediately compare your salary with theirs, because the AI retrieves this information as you have access to that folder according to the permissions structure.”

“Many scans I see around me only cover the technical side of the story”

“In addition, we also look at the human side: what is your organisation’s policy on this, how willing are you to use AI, and how far have you got with training? Because, as we just said, it’s very important that you monitor the output, that you know what it does and how it works. We explore this through an interview with the contact person, which allows us to assess what kind of support the organisation needs. Many other assessments I see around me tend to focus almost exclusively on the technical side of things.”

2. AI in our existing services

Whilst Copilot is a visible AI tool for our users, AI is also ‘hidden’ within many of the services we offer. In various IT and security solutions, AI or machine learning is used in the background to, for example, detect anomalies, analyse large amounts of data and respond more quickly to threats. AI even plays a role in user support, though the human factor remains slightly more important there.

MXDR

For our MXDR service, which guarantees 24/7 monitoring of your IT environment, we work closely with cybersecurity specialist Masero. Their platform is based on Microsoft security solutions and, via Microsoft Sentinel, utilises AI and machine learning to analyse large volumes of data and identify suspicious activity. This enhances the accuracy of threat detection and helps to reduce so-called false positives. Threats can be automatically blocked, but staff at Masero’s Security Operations Centre (SOC) always review incidents themselves first. This is because automatic blocking could inadvertently have undesirable consequences for a user. In our article ‘Every minute counts in the face of a cyber threat’, you can read more about why human assessment is so important.

AI in other security services

Most of our other services within our security levels/packages are also based on Microsoft. AI comes into play with some of these. According to engineer Matthijs Tuenter, this is certainly the case for the services Data Classification and Labelling,Risk Management for Cloud Applications, Risk Management for Internal Threats, andDetection of High-Risk Logins. Matthijs: “But it’s not entirely clear exactly what Microsoft is doing with AI behind the scenes. For us, that’s a given.”
Image

“I use a lot of AI to draft ‘fake phishing emails’”

“To provide the ‘Set up and check DNS records for email’ service, we work with DMARK AI: their platform makes extensive use of AI and features an AI chatbot that makes it easier for me to retrieve the results and clarifies them. This helps me to produce a clear report for the client.”

AI also plays an important role in our security awareness programme. Matthijs: “I use a lot of AI to create ‘fake phishing emails’ that we send to employees on behalf of a client to see how alert they are. AI helps me to get an idea of what type of email might be of interest to the company for which I’m running the programme. I think AI is also used for the programme’s standard emails and training videos.”

Using AI for phishing emails… attackers do that too, by the way. We recently held a webinar on this topic, which you can watch again:▶️ The impact of AI on cybersecurity

IT management

According to IT managers Andy Hendriks and Sebastian Setz , the management tools they currently use do not yet incorporate AI, but this will change with the introduction of Autotask. Operations Director Pierre Struijk is currently in the midst of implementing this: “Once this has been fully integrated into our organisation, we’ll be looking at our RMM (Remote Monitoring and Management) tools,which integrate with Autotask. And there’s a whole lot of AI intelligence behind that.”

The Service Desk and AI: help with queries

Our Service Desk uses AI much more than it did a year ago. Tickets are resolved more quickly because staff can find answers to our customers’ queries more easily. Service Desk agent Paul Langen uses the web version of Copilot as a support tool: “It’s like a smarter version of Google where you can simply type in a question or paste an error code, and you can even have a bit of a chat with it. Google does give you an answer, but having to read through all those pages… Copilot does that for you. It saves a lot of time. I know one of my colleagues uses Claude, as he has a licence from the school, whilst another colleague uses ChatGPT – all sorts, really. That’s because we don’t have any Copilot licences. One colleague does have a paid licence, though, so he can do more. That would help us too.”

Team leader Jenna Spelten hopes those paid licences will be available soon: “Then we’ll also be able to utilise internal data, and as you build up your knowledge, you’ll be able to provide faster and better answers to questions that Copilot has already answered before. With the free version and other free AI tools, we follow the rule that we never enter names, company names or sensitive information. With a paid environment, we can do so safely.”

Ticket communication in clear language

Jenna: “Jira, our ticketing system, also includes an AI tool called Rovo. This allows service desk staff to correct their spelling or rephrase things more clearly directly within the ticket. You see, the Service Desk uses a lot of IT terms that users aren’t necessarily familiar with. This ensures that everyone understands what we’re talking about and helps prevent miscommunication.”

“AI provides logical answers, but sometimes you don’t want a logical answer”

Real staff members rather than a chatbot

Paul: “Our users would rather speak to one of our staff than use a chatbot. I feel the same way myself; it feels as though someone is actually listening to you. And with a chatbot, you have to work through a whole list, and if your question doesn’t quite fit in, you don’t get very far and you end up feeling frustrated.” Jenna: “I agree. Empathy is essential, and that’s what’s missing with a chatbot. We can let staff let off some steam over the phone, and that’s certainly important when someone’s emotions are running high. AI provides logical answers, but sometimes you don’t want a logical answer. And it’s worth noting: a chatbot also means our staff have to keep on top of that, on top of the tickets, emails and phone calls. A chatbot with an AI function is possible, but you have to make that clear in advance and always offer the option for them to ring and speak to one of our staff members.” Paul: “I see it as a sort of FAQ, for some of the questions.”

3. Smarter apps and processes with Clevir Solutions

The handy apps we offer, such as the HR app, the Asset Management app and the Visitor app, have been developed by Clevir Solutions, part of the IT Local holding company. In the development process, our colleagues at Clevir are making increasing use of AI, and the first app featuring an AI component is now a reality.

“The Competence app is our first app with AI integration”

Skills app with AI integration

The Competence App is the first app developed by Clevir to feature AI integration. In this app, HR professionals can now automatically generate job profiles. Developer Robert Vrooland explains how it works using an example where someone from HR wants to add a new role to the existing roles within the organisation: “The employee enters the job title, the department to which the role belongs, and types in some additional information about what the role entails. The app then uses AI to search for relevant public information on the internet about this role, checks the existing internal list of competencies and suggests suitable and any missing competencies. As an HR officer, you check the result, tick the competencies you want, click ‘save’ and… your role with its job profile is ready. This applies to junior, mid-level and senior staff alike.” According to Robert, this saves HR a lot of time. “And job profiles also become more consistent because the AI also analyses the data from existing competencies.”

Other applications

Within the Clevir team, there are also further ideas for AI applications. Robert : “The next step is to try adding an agent to an app. The HR app would be a logical choice. After all, employees can have a great many questions when it comes to HR. It would then be very easy to let AI do that work, rather than having to trawl through all the HR documents and manuals and scour the internet. But it does have to be secure, of course. If I ask how many days’ holiday my colleague has left, I shouldn’t be given an answer to that.”

Here, too, it becomes clear once again how important it is to set up the legal structure properly, as Bart mentioned earlier in this article. Robert: “In Copilot Studio, which we use, we can also already catch a few things.”

Robert: “An agent like that can also initiate processes or workflows itself, rather than just providing a response. I’d like to look into that as well.”

“Integrating Copilot Chat into the workplace portal”

Fellow developer Terrence Wanga also has a concrete idea for AI integration: “Integrating Copilot Chat into the workplace portal we offer. And having the option to use agents as well. This is still in the research phase.”

Inspiration from the EPPC

In June, Terrence attended the European Power Platform Conference (EPPC) in Copenhagen. There, too, he gained new insights, particularly regarding the use of AI in the app development process. Terrence: “I foundMCP (Model Context Protocol)servers the most interesting topic. They allow you to give AI agents access to all sorts of external and internal information, making it easier for the AI to understand. You can build your own MCP server for different APIs. I’m currently testing this out.”
From Copilot to smart apps: AI within IT Local

“What was also interesting were the vibecode apps (coding based on prompts, ed.) and options in Power Automate to run workflows faster. We’re not using those vibecode apps at the moment, but I’m already testing them myself. Ultimately, the EPPC also confirmed that my knowledge is sound, as I came across many things I already knew.”

“In the past, you had to spend loads of time on UI and UX for apps; now it can be done in just a few days”

Saving time through AI in UX/UI design

Developer Sunny Jansen saves a lot of time by using AI for UI/UX design, which ensures that apps look good and function logically and smoothly. Sunny: “In the past, you’d use Google to find out more about a problem or to get inspiration. When I first started out here, I realised that with apps, you have to put a huge amount of time and effort into UI and UX – particularly colouring, configuration and checking that everything scales properly across different screens. But with the advent of AI, you can get all this sorted quickly and effectively; in a matter of days rather than spending weeks on it. You can ask AI to use a specific image, take that as a basis, and build the rest from there. It works really well, functionally too. Of course, you still need to check that it’s correct and tweak things here and there, but it saves so much work. I’m currently doing this in ChatGPT, as I don’t have a paid licence for Copilot. I’d like to get one and give it a go.”

“This approach can also be very helpful for quickly customising a client’s app welcome screen, based on a template, to reflect the client’s correct corporate identity. That’s something I’m thinking about too.”

Also, the development of AI tools by IT Local engineers and consultants

In addition to Clevir’s developers, IT Local’s engineers and consultants also develop AI tools. Bart Hunik: “At Clevir and other development firms, there are actual developers; at an IT company like IT Local, there are people who need to be able to automate certain tasks – that’s often the job of a consultant or engineer rather than a developer. If it becomes too specialised, you turn to developers. We’re tapping into the low-hanging fruit.”

Manage everything via a single chatbot

Bart: “What we’re focusing on more are chatbots. So that an AI agent can more easily resolve certain issues or operate applications for you. That’s definitely where the future is heading. As a user, you currently have a number of applications you work with, and you always have to click your way through an interface for each app and carry out various actions. Often, by asking a chatbot a specific question, you can get that chatbot to do it for you in one go. For example, if you’re applying for leave. Usually, you’d have to log in to a portal, select the relevant day, enter a reason and then submit the request. With the chatbot, it’s quicker: you state what you want, the chatbot asks for any missing information and carries out your request straight away.”

“There is potential for gain in this type of intelligence. So that more applications can integrate with such a chatbot and you can manage everything from a single chatbot. The design of a chatbot is fairly generic; the questions the chatbot needs to answer are something the organisation must provide. Once an organisation has a chatbot like this, it can easily expand it as software packages make more agents available.”

4. How do we use AI ourselves?

In the previous chapters, you’ve already seen a few examples of staff members: AI also helps us to work smarter and more efficiently. Some employees use AI tools on a daily basis, whilst others only use them for specific tasks. As usage is increasing across the board, we are also working on clearer guidelines regarding safe and responsible use.

Practise what you preach: Copilot

One thing is clear: at IT Local, we mainly work with Copilot. So, we practise what we preach. Our staff members with a paid licence who spoke in this article are frequent users: for instance, Gaby uses Copilot 99 per cent of the time: “Every day, for generating and improving text, summarising documents, and drafting manuals and plans.”

Bart also uses Copilot the most: Bart: “And in my case, I’m also talking about Copilot Studio for developing my own agents. Personally, I did find it a bit tricky at first to use Copilot directly from within the Office applications, but I’ve noticed I’m doing that more often now. I mainly use it for checking quotations. If I’m unsure about certain bits of text, it’s easy to ask within the application itself: ‘Check this text against other quotations and see if there are any discrepancies.’ I also use it to have a bit of text rewritten straight away in Word or to create a neat little table from a dataset.”

“For me, Copilot is an extra pair of eyes, a sort of extra colleague”

Matthijs keeps the Copilot chat open almost all day: “To help find a way forward for any issues that arise. Also to bounce ideas off when I need to draw up a proposal for a client regarding a project. Have I thought of everything? Are there any other points I need to include? Have I interpreted things correctly? For me, Copilot is an extra pair of eyes, a sort of extra colleague… what’s in a name…”

The Clevir and Service Desk staff who spoke indicated that they also work with Copilot, but mentioned the use of other tools within their department, such as ChatGPT, Claude and Gemini. This is mainly due to the lack of paid Copilot licences. The staff certainly recognise the added value of such a paid licence. Jenna had already mentioned this from the Service Desk’s perspective, and Robert also sees the added value for the Clevir team: “Those paid licences will become more important to us as we start working on the various new possibilities we’ve mentioned.”

A paid licence for all employees

So why doesn’t everyone at IT Local have a paid Copilot licence? Operations Director Pierre Struijk explains: “Those licences are quite pricey, so we’ve initially made them available to our staff who would use them the most. At the moment, it’s mainly our consultants and engineers who have a paid licence, because they need that extra intelligence and, for example, want to understand how things work at the client’s site and search through internal documents. However, we do want to move towards providing a paid licence for all staff in the near future. We want to encourage the use of AI – and preferably Copilot – whilst further tightening up security measures. This requires appropriate licences, as well as some training and guidelines on how to use the tool.”

IT Local – Pierre Struijk

“Working safely with AI is also of great importance for our ISO certification”

Clear guidelines for the safe use of AI

The use of AI within IT Local has grown faster than the formal agreements surrounding it. That is why we are working on a clearer framework for issues such as data protection, verification of results and the use of various AI tools. Pierre: “We are currently developing a policy with clear guidelines for our staff on how to use AI safely. Everyone is familiar with the dos and don’ts and applies them – such as anonymisation in prompting – but these have not yet been formalised. This is of great importance for our ISO certification. In the guidelines, for example, we’ll stipulate that staff will soon only be allowed to use approved AI tools, that they must handle company, customer and personal data with care, that they must always check the output, and that they remain personally responsible for the final result. And when it comes to sensitive data, our rule is: only share it in environments suitable for that purpose. If in doubt, don’t enter it.”

Copilot readiness scan: for ourselves as well

“As a good starting point, we ran the Copilot readiness scan on ourselves. The result: our foundation was sound, but to roll it out widely, we still need to make some adjustments in the area of data security in certain areas, such as organisation-wide links and external access to OneDrive and SharePoint. The scan also highlighted areas for attention regarding adoption, but we’re already working on these through the implementation of guidelines and by further promoting AI within the organisation. The overall recommendation to introduce Copilot in phases is actually exactly what we’re already doing.”

Want to find out more or get started?

Through this article, we have given you more insight into what we do in the field of AI: how we help our clients using AI and how we use it ourselves. The main message we want to convey is that AI can genuinely help you work smarter and more efficiently, but that you do need to know what you’re doing, particularly from a security perspective. We’d be happy to help you get started. For example, with our Copilot readiness scan, so you know what’s needed to get off to a safe start and deploy Copilot effectively within your organisation.

Business consultant Chris Melchers is keen to add : “Isee that many organisations recognise the efficiency gains and time savings and want to start using AI, but often haven’t properly thought through the business case behind the application. Can an AI agent really do what you want it to? How will the solution change the way you work? And what does it mean for staff? If you want a solution with broad support, you need to take a holistic view of an AI issue. We can help you with that too.”

Want to find out more or get started? Give us a call or send us an email!

Sharing? Yes please!