NGO / Social Impact
AI Project Scoping & Volunteer Matching Platform
Conversational AI for nonprofit operations, from scoping through to volunteer matching.
- RAG
- LangGraph
- Conversational AI
- CRM Integration
5x
faster volunteer onboarding, completed in minutes rather than days
90%
improvement in lead quality through AI pre-screening
75%
reduction in unnecessary scoping calls
85%
time saved in initial screening
- Personalised AI matching scales volunteer-to-project fit without proportional headcount.
Onboarding speed is the metric that governs volunteer retention — most attrition in volunteer programmes happens in the wait between offer and start.
Volunteer-driven organisations lose a large share of the goodwill offered to them in the gap between someone volunteering and someone starting. A project arrives described in whatever terms its originator used, the skills it actually needs have to be inferred, and matching it against a pool of volunteers with self-described abilities is manual work performed by coordinators who are themselves usually volunteers.
Because that capacity is limited, the delay between offer and start is measured in weeks. Volunteer enthusiasm does not survive it well, and the organisation loses people it had already successfully recruited.
Why it is hard
Neither side of the match is structured. A project description written by the person who needs help is not a requirements specification, and a volunteer's account of what they can do is not a skills taxonomy. Both are prose, written by people with no reason to use consistent vocabulary.
Conventional software responds by imposing structure — dropdowns, taxonomies, required fields — which shifts the burden onto the very people the organisation is trying not to deter. Every field added to a volunteer sign-up form costs sign-ups, and every taxonomy imposed on a project description makes it harder to describe an unusual project.
What we built
The platform handles scoping conversationally: someone describes what they need in their own words and the system asks the follow-up questions a coordinator would, producing a structured scope from an unstructured description without making the person fill in a form.
That structured scope then drives matching against the volunteer pool, carrying a project from initial description through to matched volunteers in one continuous flow rather than a sequence of queues waiting on coordinator capacity.
- Conversational scoping that asks the follow-up questions a coordinator would ask
- Structured project scope derived from an unstructured description
- Matching against the volunteer pool driven by that derived scope
- One continuous flow from description to matched volunteers, rather than staged queues
- Coordinator capacity spent on judgement rather than on intake and triage
Questions about this engagement.
Does the system decide who gets matched to a project?
It proposes matches with the reasoning behind each one; coordinators decide. Matching people to work involves considerations that are not in any profile — availability, development, who has worked together before — so the system narrows a large pool to a sensible shortlist and leaves the decision where it belongs.
Why conversational rather than a structured form?
Because form length and completion rate trade directly against each other, and both sides of this problem are volunteers who can leave at any point. A conversation gets the same structured information out while asking only the questions a particular case actually needs, which is considerably less work for the person answering than a form built for every case.
Does this apply outside the nonprofit sector?
The pattern does. Deriving structure from an unstructured description and using it to match against a pool is the same shape as internal staffing, partner matching and intake triage in general. What is specific here is the constraint that neither side can be asked to do much work, which is what pushes the design towards conversation.
What this was built on.
Tell us what you're trying to solve.
Describe the workflow, the data behind it, and what a good outcome looks like. We read every message and reply personally.