New Feature

Automation for Nonprofits: Building Confidence, Not Just Workflows

Nonprofit teams are chronically under-resourced. We built a no-code automation system from scratch, and the hardest design problem turned out to be fear, not complexity.

Role
Product Design Lead
Timeline
2022
Team
CTO, Product Manager, SDR
Tools
Figma, Linear, MailChimp, Hubspot
Automation for Nonprofits: Building Confidence, Not Just Workflows

Context

Nonprofit staff wear too many hats. Repetitive tasks (thank-you emails, donor record updates, follow-up triggers) eat hours that should go toward the actual work.

Automation was the #1 requested feature category and a growing competitive gap. I partnered with the Product Manager and CTO to frame this as more than a feature: a platform capability that would unlock future surface area for integrations, AI-assisted workflows, and custom donor journeys.

Automation and Workflow feature overview

The Research Surprise

I talked to 15 nonprofit operations managers and expected to hear about complexity concerns. Instead, the #1 fear was much simpler:

"What if I accidentally email all 5,000 of our donors?"

That insight reshaped the entire design. The problem wasn't complexity. It was confidence. People needed to trust the system before they'd use it.

I also studied the competitive landscape by testing Hubspot and Mailchimp directly using Keela's internal accounts, finding strengths and weaknesses, not just feature lists. I analyzed Phrasee, Klaviyo, Simplero, and ActiveCampaign. The research surfaced user stories we hadn't initially captured and exposed real gaps in competitor approaches around re-enrollment logic and contact-level control.

User stories and feature scope for Keela Automation
Required user stories developed with the PM. Some moved to the V2 backlog to hit the MVP timeline

The Design

Visual Workflow Builder

A node-based canvas that makes automation logic tangible. Users see their automation as a visual flow (trigger, conditions, actions) instead of filling out abstract forms. The builder deliberately constrains complexity: start from a template, extend it. No blank canvas anxiety.

Preview Mode (The Key Feature)

Before activating any automation, users can preview exactly which contacts it would affect. This was the feature that gave people the courage to actually turn things on. They could verify without risk. This design decision directly traces back to the 5,000-donor fear uncovered in research.

Final automation workflow builder — node-based canvas with triggers, conditions, and actions
The final product: visual workflow builder with branch logic and preview mode

Templated Journeys

Rather than a blank canvas, users start from pre-built journey types: New Donor Journey, Recurring Donor Journey, Disengaged Donor Journey. Each template suggests trigger timing and email structure: not the copy, but the scaffolding. Getting the scaffolding right is what most organizations struggle with.

Progressive Complexity

New users see a curated template library. As they build confidence, the interface reveals more: custom triggers, conditional logic, multiple branches. The system grows with the user.

Testing

I ran three rounds of prototype testing, each focused on a different part of the experience. The PM and Lead Designer joined sessions to record notes. Each round fed back into Linear as prioritized action items.

Prototype 1: Creating and activating a workflow
Round 1: Creating and activating a workflow
Prototype 2: Monitoring an active workflow and tracking contacts
Round 2: Monitoring an activated workflow
Prototype 3: Pausing and reactivating individual contacts in a workflow
Round 3: Pausing and reactivating contacts

What Happened

Automation launched as a premium paid add-on:

  • 20% of newly onboarded organizations purchased the add-on since release
  • Keela's MRR increased by an average of 30% monthly from the add-on alone
Keela customer case study featuring automation — Love Quilt Project
A customer case study featuring automation at the Love Quilt Project

The preview mode insight (building for confidence, not just capability) is something I keep returning to. It applies to design systems, data integrations, and any feature where users fear irreversible consequences.