Everyone talks about AI. But what does it actually get you in a cultural organisation or non-profit? And what does it cost?
The answer is simpler than you would think. AI works best on your own numbers. And you already have them: your members, your donors, your events, your grants, all in Odoo. AI can do something with that today. Not someday. Now.
Below are four things AI already does. When to start, and when to wait a bit. And why custom work costs far less than it used to.
Four things AI already does
No far-off future. Four things you can measure, running on numbers you already have.
| What AI does | How it works | What you get |
|---|---|---|
| Compile reporting | Builds reports and exports per grant and budget. | Less time on reporting, more on your mission. |
| Sort applications | Reads incoming email and applications and files them right. | Faster handling, less searching. |
| Predict lapses | Spots which donors or members are about to stop. | Reach out in time, fewer cancellations. |
| Estimate sales | Forecasts ticket sales per event. | Better planning and capacity. |
For none of these four do you have to track anything new. It is already in Odoo.
The questions we hear most
Organisations almost never call us saying “we want something with AI”. They call with a problem they are sick of. Three examples:
“Can you make our grant reporting easier?” Every euro has to be accounted for to funders and auditors, and that eats hours. With our tool togrant.com a grant becomes a central record, linked to donor, programme and budget. AI compiles the reporting and exports from that faster.
“Can AI predict which donors are about to lapse?” A donor rarely stops overnight, there is behaviour leading up to it. AI sees that in your own data and flags it, so you can reach out before they are gone.
“Can an application file itself in the right place?” Donations, members, ticketing and grants quickly end up each in their own corner. AI reads an incoming application and files it right. That is how Hart Haarlem runs events, ticketing, rental and a webshop on one Odoo, without fragmentation.
When yes, and when not yet
Not everything in a cultural organisation suits AI. The line is usually easy to see:
| Start with AI when… | Wait a bit when… |
|---|---|
| …your Odoo already thinks in donors and budgets. | …everything still sits in separate tools and folders. |
| …it is something that happens often (reporting, applications, email). | …it only happens now and then. |
| …you can measure the result in time, donors or turnout. | …nobody can say what a good result is. |
| …members, donations and accounting are in one system. | …your key numbers are still loose in spreadsheets. |
Put AI on a shaky process and you mostly get faster mistakes. Get the basics right first. Then AI.
Why custom is only affordable now
A non-profit always used to keep two lists. One had everything the software should do. The other had what it was allowed to cost, and on a tight project budget especially. The good ideas got stuck in between. That gap has largely closed.
The reason is simple: building goes much faster with AI in the loop. People call it vibe coding. The work shifts. Less typing code, more thinking clearly about what you want, then checking it is right. That is where the craft sits.
An example. togrant.com turns a grant into a central record, with donor-structured budgets and ready-made exports like the EU PRAG report. Building something like that used to be too expensive for a non-profit. Not anymore.
That is the real change. Not that custom got a bit cheaper. But that a whole pile of “nice, but too expensive” ideas suddenly became possible.
Where to start
Start small. Pick one job that hurts right now and where your data already lines up. Usually that is reporting or sorting applications. Prove it there first. Works? Then the rest. And keep your Odoo in order, because that is what AI stands on.
Still weighing up whether Odoo fits your organisation? Start with Odoo for culture and non-profits and why grant management is the key. And read why custom is affordable for mid-market now for the full story.