Key insights
- Licences are just one part of your AI budget.
- Get your data and rights in order before you roll out your AI.
- Agents and AI services are often charged on a per-use basis.
- Invest in adoption and training: that’s where the greatest returns lie.
An AI budget is more than just licences
When it comes to an AI budget, many organisations first think of licences, such as Microsoft 365 Copilot. That makes sense, as that is the most obvious cost. But in practice, it is mainly the other items that determine whether your investment pays off. A licence that nobody makes proper use of costs money without delivering any results.
We therefore recommend structuring your AI budget for 2027 around five categories. In our AI Implementation Programme We’ll go through those items with you.
1. Licences in your AI budget: who are they for and which ones?
Not every employee needs the same level of AI support. Identify which roles stand to save the most time with AI, and start there. Also take a look at your existing Microsoft licences: sometimes a bundle with Copilot works out cheaper than a standalone add-on. You can read more about this in our blog post on the Microsoft 365 pricing update.
- Which teams or roles start first?
- Is the free version enough? Microsoft Copilot Chat for some of your staff?
- Does your current mix of licences still fit in with your plans?
2. Implementation and data: is your environment ready?
AI works with the data that staff can already view. Are the permissions set in SharePoint If Teams isn’t set up properly, Copilot will also find information that isn’t intended for everyone. You should therefore set aside a budget to get your environment in order beforehand:
- check permissions and access;
- set sensitivity labels, for example using Microsoft Purview;
- clearing out outdated or duplicate data.
3. Usage: agents and tokens
Anyone who goes beyond Copilot in Word, Outlook or Teams will soon find themselves building their own agents or integrating AI into business processes. Examples include Microsoft Copilot Studio (see also our blog Copilot Studio: build your own AI) or AI services in Microsoft Azure. These solutions are often charged on a per-use basis: per message, per action or per token used.
This means that the costs can vary. You should therefore make a realistic estimate of consumption and agree on who will monitor it.
4. Adoption and training: get the most out of your AI budget
You should therefore set aside part of your AI budget for training and support. That way, you’ll really get the most out of your licences. On our events Incidentally, AI is often discussed, with demos and real-world examples.
5. Governance and cost management
Employees are increasingly building small AI applications themselves, such as dashboards or simple apps. This is a positive development, provided that IT retains an overview. You should therefore set out the following in advance:
- which AI tools are permitted;
- how to protect your business data;
- how to monitor your AI expenditure and usage.
This way, you can avoid any surprises on your bill and ensure that AI remains safe and manageable.
Your AI budget starts with a plan, not with a figure
A good AI budget doesn’t start with a figure, but with the question of what you want to achieve with AI. Which processes do you want to speed up? Where are staff currently wasting time? Based on those objectives, you can determine which licences, support and usage levels you really need. With us AI Implementation Programme We’ll guide you every step of the way, from strategy to roll-out.
Are you planning your AI budget for 2027? Our AI experts would be happy to work with you. During a brief introductory meeting, we’ll assess where you stand today. We’ll then plan an analysis session together to fine-tune your budget and draw up a concrete AI plan for 2027.
VAT no.
Author
Article written by
Olivier Laloo has been helping organisations build their digital capabilities since 2007, with extensive expertise in Microsoft Dynamics 365 Business Central and Finance & Operations. Within the VanRoey ICT Group, he leads the way in the field of AI: from strategy and budgeting to the practical implementation of Copilot and AI agents.
share this post: