Bridgit AI Rules: now in beta for all

We've been working hard to make bridging two platforms easier for our customers.

Ask any charity admin, getting the data is only the first step. The really time consuming work is in everything that comes after. Cleaning data, uppercase, lowercase, calculations, filters, combining, summing values - the list goes on. One platform gives you KATIE and statham in two columns; the other wants Katie Statham in one. Split donations need excluding. A size field of "child" needs to become "Child - T-shirt", and everything else needs to become "Adult".

None of it is hard. All of it is fiddly. And historically it has meant one of two things - learning an expression syntax you'll use twice a year, create an Excel function, or opening a ticket and waiting for someone who already knows it.

Today we're changing that. Bridgit AI Rules are now generally available.

Say what you mean

Open any field in your mapping and describe what it should contain, in the same words you'd use to explain it to a colleague:

"Combine the first name and last name and call it full name - make sure it's titled properly"
"If donor is a child, return 'Child - T-shirt', everyone else gets 'Adult - T-shirt'"
"Filter out split donations"

Bridgit reads that, builds the rule, and explains back what it did in plain language - "Combines the donor's first and last name into a single value, capitalising the first letter of each name and lowercasing the rest." No syntax to learn, no documentation to open.

See it work before you commit

Once the rule is generated, you can preview the data. Bridgit runs a quick API call on your behalf and collects real rows from your actual data, side by side, before and after. KATIE / statham becomes Katie Statham. A filter tells you plainly that 2 of 4 rows were kept. You see the outcome before anything is saved.

If it isn't right, refine it with another prompt to Bridgit. If you'd rather take complete control, open the advanced view and edit the expression directly - it's a readable rule, not a black box, and it stays yours.

Your data never reaches the LLM

I want to be very clear about this (and brag a little), because "AI-powered" has come to mean "your records are somebody else's training data", and for us - that wasn't good enough.

None of your data is sent to the LLM.

The model helps construct the rule. It never sees the rows the rule runs against. The preview, the transformation, and the run all happen inside Bridgit. Your donors' names, addresses, and giving history never leave the platform, not to a model provider, not to anyone but you.

The LLM builds the rule, Bridgit transforms your data.

Where it works

Bridgit AI is now available across every bridge you build.

Sources: more than 20 fundraising and payment platforms, including JustGiving, Enthuse, Donorbox, Raisely, GiveWheel, GivePanel, MuchLoved, WeGive, Fundraise Up, iDonate, Funraisin, GoDonate, GoFundMe and Stripe. We're adding more all the time.

Destinations: Spreadsheet, S3-compatible storage, and SFTP, delivered straight into your existing workflow. CRM destinations are now in beta.

Get Started

Bridgit AI is live for all customers today - no upgrade, no waitlist, no flag to request. Open a bridge, click into a mapping rule, and describe what you need.

Build your bridge, we'll handle the rest.

If you want to get started today - login your account or want to know more about how the rules engine works reach out to me at [email protected]