With many companies now working on their new multi-year plans, the question of a “digital strategy” is back on the table. We explained earlier why you don’t need one. Lately, though, a second question has been popping up: do we need an AI strategy too? So the documents keep piling up: a digital strategy, an AI strategy, a diversity plan, fair pay policy, the annual plan and new funding applications. Each comes with its own ambitions, goals and activities, sometimes even a different mission, which at best complement and complete each other, but more often than not are completely misaligned.
The question is understandable. No technology has found its way into our work as quickly as AI. Instagram needed 19 months to reach 50 million users. ChatGPT did it in five days. Some people refuse to use any form of artificial intelligence, but for most, the question is no longer whether they use it, but how.
Whereas organisations used to have time to get used to new technology and make a long-term plan, that space has all but disappeared. Almost every week we see new consequences of AI. One week we’re debating intellectual property and AI; the next, websites have to go offline because of a weakness in PHP systems that AI suddenly exposed. A plan for “next year” has become a bold undertaking.
The good news: cultural organisations are trained in improvising, adapting and making something beautiful with limited resources. What matters is anchoring that agility in your strategy too. That’s why we developed a workshop at Tò Mò in which we work with your team, in one afternoon, to develop and write down a vision on AI. Based on that experience, we share five tips here.
As a director, you want to make use of the knowledge in your team. We see that it works well when the team is involved in drawing up the vision on AI, with the leadership keeping a close eye on the organisation’s mission and vision, and drawing on the expertise in the team. If the writing of the AI vision is immediately outsourced by the leadership to an employee or a member of the management team, there’s a good chance you’ll get stuck. A vision like this is built together; you can’t just pen it down in between meetings. It’s important that the leadership sets the direction and then brings everyone along.
“Accept uncertainty. We can’t foresee where the technology will be in one, two or five years, so we can’t write plans that cover every scenario”
Bring all employees into the conversation. Start from your mission and discuss where the opportunities lie and where the limits are. Where can AI help you achieve your goals? Which organisational problems do you want to tackle, and could AI be a solution?
Talk about the ethical side too. At Tò Mò, we use as few American providers as possible and look at what Mistral can do. We also think it’s important that our texts are grounded and authentic. This blog was written by our founder Linda Thoen in collaboration with our copywriter Evelien Sinke. The human creation this requires is, in our view, something AI cannot take over. We do like to use AI to check emails for spelling, to double-check the figures in a quote or to visually straighten out presentations. For us, that is where the line lies at the moment. We may revisit it soon, which is why this conversation is never finished.
For this ongoing conversation, you lay a foundation in a first conversation or workshop. Set aside at least half a day and consider a facilitator. Such a conversation is successful when it’s clear when you’ll explore the possibilities of AI, and when you definitely won’t use it.
You might, for example, conclude that it would be worth exploring whether software systems that use AI could reduce the workload in marketing or production, and that it would not be desirable to use AI to take over artistic work processes to any degree.
In conversations, AI sometimes feels big and elusive. Often, though, we’re talking about software, such as planning software or content creation programs, that uses an AI model. Think of the AI function in Canva or in your email client. There are of course chatbots like ChatGPT, and AI is now also built into office suites like Office and Workspace.
Getting an overview of everything AI can do is impossible. So start concretely: where is the problem in your organisation? You can use a chatbot to explore what’s possible. Test a few things on a small scale, get familiar with it and stay critical. Can the company behind the model be trusted? Do I want to give this company access to certain information and processes? Which data are we using, is it our own data or dummy data, and where does the data we share with the AI model go? Always check the terms and conditions of the service carefully, and consider whether they match your wishes and boundaries.

You can’t foresee everything. Nobody had included a playbook for a PHP hack in their digital strategy. The longer the document, the less you actually pin down, and with AI that’s even more true. We don’t know where the technology will be next year. An AI vision, as part of your organisational strategy, fits on one or two pages.
Bundle your digital, diversity and AI strategies into one organisational strategy and begin with your mission. Why do you exist? What would society miss if you could no longer open your doors? If your mission is longer than two sentences, try to distill its core.
Link the mission to three to five annual goals. Make at most one or two of them SMART and keep the rest abstract.
In conclusion: how to start tomorrow
An agile vision on AI is partly put down on paper, so that everyone in the organisation can see it. It is mainly an abstract vision or ambition, preferably part of your overall organisational strategy and always linked to your mission and annual goals.<br/><br/>
How do you start? Plan half a day with your team this week, with external expert guidance if you like, and begin with one question: where do our mission and goals lie, and where can AI help us achieve them, or not? You don’t need to come out with a ready-made vision. A clear first conversation is a perfectly good start.<br/>
And accept the uncertainty. We can’t foresee where the technology will be in one, two or five years, so we can’t write plans that cover every scenario either. Write less, and make sure colleagues can decide for themselves, based on the vision, when to use AI in their daily work and when not to.
Deze website maakt gebruik van zowel functionele, (geanonimiseerde) analytische als tracking cookies. Deze gebruiken wij om de website zo goed mogelijk te laten functioneren en het gedrag van bezoekers op deze website te analyseren.