AI Search Experience — Benevity

Context

Benevity's Goodness platform lets employees search for volunteering opportunities, nonprofits and giving campaigns across a catalog of hundreds of thousands of organizations. The existing search behaved like a plain keyword lookup — broad, generic, and blind to intent.

This project explores what search could be instead: a conversational, AI-assisted experience built on semantic understanding, real-time synthesis and personalization, rather than exact string matching.

A volunteer at a community event — the world Benevity's search is meant to connect employees to.

The product lives in the community: employees looking for a volunteering shift, a cause, a place to show up.

My role

I'm leading the design exploration end-to-end — framing the problem, defining design principles for AI-assisted search, and designing the 60% proof-of-concept flows now in review with engineering and product ahead of a broader design-review cycle.

Problem to solve

Discovery surfaced five recurring pain points with the existing search experience:

Inconsistent functionality across domains — siloed, rigid, and out of step with modern user expectations.
Incomplete or irrelevant results, with little contextual understanding.
Poor search accuracy and precision.
Limited, hard-to-use filtering options.
General usability and intuitiveness issues.

Expected outcome

The goal: someone should be able to find a volunteering or giving opportunity that matches their intent within one or two searches — with results that clearly explain why they're relevant, and let people act immediately (sign up, donate, save, share).

Prioritize empathy & trust

Focus on actionable outcomes

Be specific & impact-oriented

Foster community & connection

Grounding the design in a scenario

Rather than designing against an abstract feature list, the team built a short narrative around Priya, an employee who regularly supports local animal shelters:

Priya is an animal lover. She often supports local shelters and nonprofits.

When she searches “donate to an animal shelter,” the results feel overwhelming and generic. She wants an easier way to refine.

Priya wants to see, in the search results, proof of impact: stories, outcomes, and exactly how her donation will be used.

Benevity AI supports Priya's search by understanding what inspires her to give and what the nonprofit needs most now.

Capabilities

Two capability tracks shape the experience:

NLP

Semantic search — understanding relationships between words, not just keyword matches.
Parses a query's intent: subject, cause and action.
e.g. Animal Shelters · Animals with special needs · Donation
Conversational interface that keeps refining as the user types.

Real-time data

Data synthesis and summarization.
Prioritization.
Personalization based on user and company signals.

The design

A full walkthrough of the 60% proof-of-concept — from the personalized dashboard through conversational search, AI-generated overviews and filtered results — currently in review with engineering and product.

A narrated run-through of the flow: the personalized greeting and dashboard, a conversational search, the AI-generated overview grounding the results, filtering, and results grouped by intent.

Status

Delivered a 60% proof-of-concept spanning dashboard, conversational search, filtering and AI-generated overviews.
Defined four design principles for AI-assisted, empathetic search, now referenced across the design team.
In review with engineering and product ahead of the next design-review cycle.

Reimagining search with conversational design and AI.