AI in your organisation: from purpose to application
By now, every organisation knows: we need to do something with AI. Sometimes there's only the ambition and enthusiasm to get started with AI. Perhaps the spot has even already been decided: in the search function on the website, a chatbot that can give customers a quick answer, or a tool that should help bring in interesting leads. The wish is clear: an organisation that can remain efficient and competitive. An organisation where the potential benefits of AI are explored in every department, and where innovation and improvements create value, both internally and externally.
Starting with the technology doesn't automatically lead to sustainable change
Experience teaches us that organisations often start their AI implementation from one of these starting points. Each of these starting points is valuable and necessary, but none of them immediately results in organisation-wide adoption of the technology or a uniform approach.
- Rolling out AI tool licences for everyone: Capacity is purchased, and in the short term there's experimentation, but that doesn't yet create demand within the organisation's teams. After six months, we notice big differences in adoption from team to team, and a large gap emerges between employees in terms of knowledge and enthusiasm.
- Writing an AI policy: A team that's often technically well-versed maps out a policy. That policy takes all possible risks into account and puts safety first. The policy provides a first piece of the framework, but doesn't yet help employees assess what genuinely valuable AI use looks like.
- Having a working group figure out what's possible: The objective is there, and now it's up to a few employees to make it concrete. Unfortunately, without a specific question to answer, the result is often an exhaustive inventory of possibilities, but without concrete direction.
- Launching a pilot in the department that's eager to go: A motivated team in the organisation actually sets something up that works and adds value: that's already a first success. However, the impact of this success on the AI transition of the whole organisation is limited: the rest of the organisation only learns from the knowledge gained to a limited extent, and adoption doesn't yet spread more broadly.
Each of these starting points delivers something. Teams get to know the technology, gain insight into what works, and build up the experience you need to later scale structurally. What they don't do is bring that experience to the whole organisation. Do you want to structurally raise your organisation's maturity? Then we also need to look at the context around the technology. The following three angles are a great complement to AI licences, an initial policy, an analysis of the possibilities, and a few concrete AI experiments. Three themes that ensure an organisation is truly ready to integrate AI in a valuable way into its current way of working.
How is 'the work' done today?
You can't assess what AI means for your organisation without understanding why people do what they do. What do they do, why in that particular way, and how would they ideally like to be able to do it? Above that sits the question of what the organisation wants to achieve, regardless of any technology.
That 'why' matters. Almost every way of working has a reason: a rule, a system that doesn't allow otherwise, or a habit that no one questions any more. Without asking that question, you run the risk of automating a workaround instead of tackling the root cause.
The workarounds that people have built themselves are the most valuable starting point here: the Excel file next to the official system, or the folder of templates that someone keeps privately. An employee or team freed up precious time because it was a problem worth solving.
What data is circulating? And where does it sometimes go wrong?
Here we try to map out what data is produced, where it comes from, and who uses it. When people think of information, they often only think of databases and systems. But it can just as easily be about knowledge, quotes, emails, reports, presentations, or product sheets.
It's also important to look at where things sometimes go wrong, where data or information is incomplete or doesn't even exist yet. This gives insight into the things that take up a lot of time, where colleagues do duplicate work, or have to wait on each other. It points to where the current way of working falls short. Employees themselves almost always already know where the problems lie; it just has never been mapped out this way before.
How can AI help us with this?
Only now does AI come into the picture. By placing the bottlenecks in the work alongside the data and information flows, we can make a distinction: which problems genuinely lend themselves to AI, and which are actually a process, information, or agreement problem that doesn't need any technology at all? By making smart choices here, an organisation can invest in the problems where AI really can add value.
Work out with the relevant teams what the impact of a solution could be: how much time is freed up, which errors disappear, and, perhaps most importantly, what the customer notices from it.
Not everything will, of course, already be in place to move straight to implementation. Sometimes information is missing, people lack certain skills, or agreements between teams still need to be defined. These questions are often skipped, but discussing and addressing them is crucial, since they determine the success of a project.
Why should you also invest in a broader analysis?
In our view, most of a project's impact is created before the building even starts. Which problem do we choose? With what information and expectations do we start?
This approach maps out where things go wrong, why they go wrong, and where AI can add value. Not all bottlenecks from this analysis require costly technology, but by tackling them, you create fertile ground for the adoption of change, with or without AI.
This way, we ensure that the investments you make in AI genuinely offer a solution to the root cause of a real problem, and don't create a new one.
Technological developments are moving extremely fast. The tool and technical setup you choose today may already be redundant or outdated in a year's time. Your people, your information, and your agreements will not be. Whoever invests in those today isn't just building one AI application, but laying the foundation for every solution that follows.
