Turning One-Off Donors into Regular Givers: Save the Children's AI Personalisation Pilot

Introduction

Save the Children works in more than 100 countries. Its Non Member Markets team looks after supporters in countries with no local Save the Children member organisation, which makes for an unusually mixed audience: tens of thousands of donors across many countries, currencies and languages, contacted through a single email programme.

In spring 2026, the team commissioned Giving Analytics to run a pilot based on AI-generated emails, each one written for and personalised to an individual donor. The pilot had one primary objective: convert single-gift donors into regular givers, the hardest ask in email fundraising.

The Challenge

The hardest ask in email fundraising

A regular gift is an open-ended monthly commitment, not a one-off donation. Persuading a one-time donor to make that commitment by email is a challenge for every charity, and Save the Children wanted to know whether personalisation could make a measurable difference.

Three very different audiences

The pilot had to work for three segments with different relationships to the charity:

• Active regular givers, who needed to be thanked and reassured, not asked again
• Lapsed regular givers, whose past commitment deserved recognition before any invitation to restart
• One-time donors, by far the largest group and the primary target for conversion

No new systems

Anything the pilot produced had to be sent through Save the Children’s existing ActiveCampaign platform using the standard workflow, with no CRM integration and no new tools.

The brief was to prove a replicable, transparent pipeline that other Save the Children member offices could adopt, not just to run a one-off experiment.

The Solution

Giving Analytics applied the personalised communications approach developed with Muslim Charity and UNHCR, adapted to a multi-country audience and a regular-giving objective.

From data to send

1. Data and segments: Five years of anonymised donation history, split by regular-giving status: active, lapsed and never.
2. Donor signals: Around 20 behavioural signals per donor, covering loyalty, generosity, frequency, favourite causes and community.
3. Personalised narrative: The strongest storylines were selected for each donor, and gift amounts were translated into concrete impact using an impact table completed by Save the Children.
4. AI generation: A large language model wrote each short email from the donor’s verified facts, with segment-specific tone and call-to-action rules.
5. Quality assurance: Proprietary guardrails and a custom testing suite checked every email for accurate names, consent, dates and claims. Save the Children reviewed samples over two feedback rounds before full generation.
6. Delivery: A send-ready spreadsheet was mail-merged through ActiveCampaign as normal.

Results

The pilot was measured against every ask campaign sent to the same segments in the 12 months before it, using standard statistical tests to confirm each uplift was too large to be chance.

Single donors became regular givers

• 23.9x the single-gift to regular-giving conversion rate of the previous 12 months of ask campaigns
• The comparison year included the charity’s strongest giving period, the December year-end series, so the baseline was not flattered by season

Engagement rose in every segment

Clicks and list health

• Around a third of all pilot clicks became a donation or regular-giving signup, close to the rate of dedicated emergency appeals, with no urgency framing at all
• Active regular givers responded to a thank-you with only a gentle donation link by giving additional one-off gifts at a rate above the norm for dedicated ask campaigns
• Unsubscribe rates stayed at or below the ask-campaign norm, so the gains came at no deliverability cost

Key Learnings

1. The hardest ask works: Personalisation converted single-gift donors to monthly giving at many times the baseline rate.
2. Simple subjects outperform: A plain “Thank you” beat every emergency subject line of the previous year on opens.
3. Soft asks convert better: With no urgency devices, one in three clicks still converted, and far more of them into regular gifts.
4. Stewardship pays: A thank-you with one gentle link drew unsolicited gifts from committed donors.

Conclusion

The pilot was recommended to scale to the full eligible audience on a regular cadence. The personalisation engine works in any market, because segmentation, storylines and impact statements are settings rather than code, and multi-language delivery had already been proven with UNHCR in Spanish and Arabic. The only data requirements are a donation-history file and an impact table, with no CRM integration needed.

Transform your donor communications

Ready to convert more of your one-time donors into regular givers? Learn more about our personalised communications service or get in touch to discuss your campaign.