Missed follow-ups in donor communication
A nonprofit organization contacts around 90,000 people a month. Responses are scattered, and follow-ups are planned by hand.
- Nonprofit
- Marketing
Key metric
Qualified follow-up contacts per month
1,2002,880contacts
Qualified follow-up contacts per month before and after introduction
At a glance
Starting point
A nonprofit organization active across Germany looks after around 280,000 active donors. About 15 campaigns run each month across email, social media and the website, reaching around 90,000 contacts in total. Responses come in as clicks, form inquiries, small donations and questions by email. Every week, the marketing team exports lists from the CRM, the mailing tool and web analytics, matches them in spreadsheets and manually marks which people are candidates for a call, a personal email or another mailing. Realistically, only about 80 hours a month are available for this review.
- 90,000
- Contacts per month
- 15
- Campaigns per month
The challenge
Because the data is scattered and time is short, the team focuses on a few obvious signals, such as large donations or direct questions. Many smaller but relevant responses go unused, for example repeated clicks without a donation or abandoned forms. Manual prioritization means that only around 1,200 contacts a month are actively followed up, even though about 4,000 people show behavior that suggests a conversation or a personalized message would be worthwhile. At the same time, data protection rules and communication preferences have to be respected, which requires additional checks and delays follow-ups further.
Typical symptoms in the company
Why this remained unsolved
The problem was too big for purely manual work but too fragmented for a classic IT project. Earlier attempts with simple scoring in the CRM failed because they only looked at individual channels and did not include free-text signals such as emails or form requests. Full automation was ruled out for fear of inappropriate outreach, since the team still has to take responsibility for every contact. So the team stuck with laborious Excel lists that could cover only part of the relevant contacts.
Our solution
The organization introduced an agentic workflow that bundles all relevant signals from the mailing tool, CRM, web forms and email inboxes every day. The workflow analyzes click paths, donation amounts, response frequency and free-text requests and turns them into suggestion lists for follow-ups. For each person it generates a suggestion type, such as a personal call, an individual email or an invitation to an online event, with a short rationale and a priority. The marketing team works through these suggestions in a simple dashboard, can adjust or discard them and gives feedback that flows into further prioritization.
Where generative AI comes in
Generative AI evaluates free text from forms and emails, summarizes requests and assigns them to follow-up categories. It suggests subject lines and conversation openers, which the team adapts. Unlike rigid rules, it understands different spellings, mixed motives and multilingual responses, and so turns weak signals into structured tasks.
- Daily merging of response data from the mailing tool, CRM, web forms and selected email inboxes into a single view.
- Agentic workflow that evaluates signals per person, assigns priorities and generates concrete follow-up suggestions with a rationale.
- Generative analysis of free-text requests to identify motives, topic interests and urgency across all channels.
- Workspace for marketing and fundraising where suggestions are filtered, assigned, documented and given feedback.
- Rule set that takes opt-ins, suppression lists and contact frequencies into account and automatically hides unsuitable suggestions.
Rollout in clear steps
Process mapping
Together with marketing and fundraising, campaign types, follow-up logic, data sources and data protection requirements are captured in detail.
Data connection
CRM, mailing tool, web forms and selected email inboxes are connected, signals standardized and test data provided.
Workflow design
The agentic workflow is configured, prioritization rules and categories are defined and tested with sample data.
Pilot phase
Some campaigns run through the new dashboard; the team rates suggestions, adjusts rules and gives structured feedback.
Rollout and training
After the pilot, all relevant campaigns are switched over, staff are trained and a rule set for ongoing adjustments is agreed.
Measured: after 4 months in production
Period: after 4 months in production.
| Metric | Value | Basis |
|---|---|---|
| Qualified follow-up contacts per monthKey metric | +140% qualified contacts | System measurementmeasured |
| Processing time per contact suggestion | -55% time per suggestion | Time tracking before/afterestimated |
| Suggestions with manual correctionQuality measure | 15% need adjustment | Samplemeasured |
Cost-effectiveness
The solution depends heavily on consistent campaign tagging and well-maintained communication preferences. If channels are not connected or opt-ins are documented incompletely, potential remains untapped and additional manual checks are needed.
Integration and operations
Also fits
A similar case in your organization? Let’s talk.
Briefly describe your process. We’ll get back to you with an initial assessment of whether the pattern can be transferred, and what it realistically takes.