USE CASES
Automating appeal coding with AI
Automating appeal coding with AI
How Wales Air Ambulance automated GoFundMe appeal coding with 98% accuracy
Wales Air Ambulance now codes its GoFundMe fundraising pages to the right appeal and campaign automatically, with around 98% accuracy, using rules it wrote in plain English with Bridgit's AI Transformation Builder.
| Charity | Wales Air Ambulance Charity |
| Platforms connected to Bridgit | GoFundMe, JustGiving, Enthuse, MuchLoved and Raisely |
| CRM | Raiser's Edge NXT |
| Focus | Automatic appeal and campaign coding for GoFundMe |
| Result | Around 98% coding accuracy |
About Wales Air Ambulance
Wales Air Ambulance is the only air ambulance charity dedicated to the people of Wales, bringing advanced critical care to patients across the whole country 24 hours a day by helicopter and rapid response vehicle. It is funded entirely by donations, raising £13 million every year to keep the service running. With a team of around 100 staff and no in-house developers, the charity relies on tools its fundraising and data teams can run themselves.
The challenge
GoFundMe is one of the most popular ways for supporters to raise money for Wales Air Ambulance. People use it to tell their own story: a sponsored run, a birthday, a tribute to someone they've lost, or a thank you to the crew who saved a loved one. That variety is exactly what made the data hard to handle.
Gifts from a single platform could belong to many different appeals. Coding each one correctly needed someone who knew every appeal, every campaign and all their nuances. For years the team monitored new pages, created supporter records by hand in Raiser's Edge and kept fundraising totals in notes on the record.
That worked, until the charity wanted to introduce propensity modelling, with our friends at Dataro, and a new suite of digital marketing tools. Supporters fundraising on third-party platforms sat outside the reach of both.
As Laura Coyne, Fundraising Operations Manager at Wales Air Ambulance, explains: "Getting third-party data into Raiser's Edge properly took about half a day of specialist setup per platform, and only one (superstar) on the team had those skills. So platforms like GoFundMe, where gifts land in lots of different appeals, stayed manual."
The solution
Wales Air Ambulance connected GoFundMe to Bridgit, alongside JustGiving, Enthuse, MuchLoved and Raisely. The breakthrough came from Bridgit's AI Transformation Builder.
Laura already had the knowledge written down: a 12-page document setting out how the charity codes third-party fundraising pages, with nine rules evaluated in priority order. She explained those rules to Bridgit's AI in plain language, much as she would brief a new colleague. Bridgit's AI turned them into a precise, sophisticated rule that reads each page's title and description and returns the right appeal and campaign.
Laura describes it simply: "I spoke to Bridgit's AI like a colleague I was training, explaining the rules for each appeal. No coding required, because I cannot code."
The rules handle the nuances that defeat simple keyword matching:
| What the fundraiser writes | How it's coded | Why |
|---|---|---|
| "Donations in memory of Mum instead of flowers" | In Memory Donations | Passive giving in someone's memory |
| "We're running in memory of Dad" | In aid of | A sponsored activity beats in-memory language |
| "Donations instead of presents for my birthday" | Celebration | Gifts marking an occasion |
| "For my 50th birthday I'm climbing Snowdon" | In aid of | An activity beats a birthday |
| "We're holding a party for Mum's birthday" | Celebration | Parties stay Celebration, even though they're events |
| "I'm running the Cardiff Half" | Cardiff Half | A named event always comes first |
Good to know: Bridgit's AI Transformation Builder never reads your supporter records. You describe your rules in plain English and the AI turns them into a sophisticated piece of logic. From then on, Bridgit's own system applies that logic to every record in exactly the same way, so results are consistent and predictable, supporter data stays private and running costs stay low.
Bridgit's AI gave Laura a working first draft, so she never had to start from a blank canvas. She tested it, adjusted it where it got things wrong and added key Welsh words for supporters who write in Welsh. She also set a fallback appeal for each platform, so nothing is left uncoded.
As Laura puts it: "Bridgit's AI drafted the appeal coding rules and I refined them, rather than having to start from scratch."
Going further: identifying service users
Many GoFundMe pages exist because Wales Air Ambulance cared for the fundraiser or someone close to them. There's no checkbox for that, but it's often in the story. Laura created a second field that looks for phrases such as "I was rescued by" and flags the supporter as a likely service user, meaning someone the charity's crews have cared for or their family. That way the charity can make sure its communications are appropriate and sensitive.
In Laura's words: "People use GoFundMe to tell their story in their own words. With Bridgit, I can have a field read a fundraising page's description, look for phrases like 'I was rescued by' [one of our helicopter crews] and identify them as a service user. This important context just wasn't findable before Bridgit."
The results
The GoFundMe bridge runs on a schedule. When it does, the team downloads a CRM-ready spreadsheet from Bridgit, with every field formatted and every appeal and campaign coded, and imports it into Raiser's Edge NXT.
- Around 98% of pages coded to the correct appeal and campaign
- GoFundMe data checked against the PayPal Giving Fund payout before going live, and "it all tallied wonderfully!" according to Laura
- GoFundMe supporters are now within reach of the charity's propensity modelling and digital marketing tools
- Supporters who were cared for by the charity, or whose families were, can be recognised and treated with care
For Laura, it's a big step forward: "We're finally at a place where we might achieve what fundraisers call the golden record: a 360 view of a supporter's interests, motivations, activity and giving history."
What's next: straight into Raiser's Edge NXT
Data from Raisely already flows directly into Raiser's Edge NXT through Bridgit, with no spreadsheet step at all. The next stage is to bring GoFundMe and the charity's other platforms onto the same direct route, so coded data lands in the CRM automatically.
Laura sums up the reaction inside the charity: "Honestly, you are the saviour of the charity. Everyone's asking how we've suddenly solved this, and I just say: 'Bridgit'."
Book a call to see how Bridgit can automate appeal coding for your charity.