The Practicalities of AI
Deniz Hassan
11th December
Over the summer I embarked on a mini European speaking tour on how frontline fundraisers can use AI right now to make their jobs more effective. I'm lucky enough to look across the full fundraising ecosystem - from strategy to tech to tactics - and I've found uses across the lot. To put it into context, I’m a digital fundraising expert who uses AI frequently and having spent a lot of time thinking and testing, here are a bunch of things I've learned using it:
It's a baby. A very clever baby but at the start of its useful life nonetheless. It gets stuff right but also gets a lot of other stuff wrong. So we can't just expect perfect outcomes.
It tends to get stuff very wrong because we haven't fed it correctly. In much the same way as people, you get out what you put in. If we're poor with how we communicate with it, the margin for error is pretty big. We make basic assumptions with a lot of the stuff we do - and with AI, this can lead to the weird and the wonderful (/bloody awful).
Both us and the AI need to learn together so we're not always starting from scratch. If, while you're working on something you make a mistake and clock that there's a better way to work, note it down for next time. Don't expect it to remember what you did and nail the same task in the same way each time. On a number of occasions, I worked with Chat GPT to undertake repetitive tasks and I found I could get the same outcome 9 out of 10 times and then the 10th would be massively different. So now I'm microscopically specific each time and repeat myself.
I’ve used it extensively for heavy data analysis and found that I need to work in a lot of quality assurance. One of the tactics I've developed is asking it for granular steps so I can compare against my own workings. Only once I've done enough spot checks do I trust it to do an entire job. People asked 'what's the point if you still have to work hard?' to which I respond that, even if in the first instance it only saves me 5% of time, then I know that next time it will be more because I've learned how to drive it better. And of course, 5% is still 5%.
It's an excellent troubleshooter. I use a number of enterprise level platforms to build anything from data visualisations, CRM configurations and automations. Often, it won't take the job off my hands but it will be able to look at my work and help me get round obstacles I might otherwise get stuck on.
It's not a fundraiser but it can be taught fundraising (scarily better than many fundraisers I've met!). It doesn't get big strategic things like how you could create a diverse, balanced portfolio 'out of the box'. But I found that, given the right level of love, it will help with strategic planning and modelling.
These are just a few thoughts and I’d love to hear what you think. And of course, if anyone wants to discuss how we can help you leverage AI in your fundraising programmes, give me a shout. You can contact me at deniz@aawpartnership.com.