Make a Difference Support Services is a startup SIL provider in Western Sydney and Southeast Melbourne. NDIS Provider ID 4-LBNGAV9. Two service regions, no waitlist, no referral history. Matt runs it.
Before the agent went live, the outreach was Matt, a phone, and a Facebook feed. He had replied to 15 coordinator and family posts using the templates we first handed him. The result of those 15 replies: zero. No placements, no referral, no coordinator who remembered the name a week later.
That is the part most case studies leave out. So we are leading with it.
Coordinator hand-raises
5
9 days after the Referral Agent went live — from a standing start of zero
What "Zero" Actually Meant
A startup provider has the hardest version of the referral problem. There is no past placement a coordinator can ask about. There is no other coordinator who has worked with you. You are the unknown provider in a feed full of providers a coordinator has already been burned by.
The 0-for-15 was not a volume problem. It was a copy problem, a timing problem, and a channel problem stacked on top of each other.
The NDIS Facebook groups have a short response window. Multiple providers reply to every accommodation post within minutes. The replies that land are specific and peer-to-peer. They name the exact post, acknowledge what the coordinator or family is carrying, and read like a resource rather than a pitch. The templates Matt had been sending were generic, too structured for the channel, and often too slow. By the time the reply went up, the post had aged out of relevance.
The honest diagnosis
Fifteen replies, zero results, is not bad luck. It is feedback. The channel was right. The contact list was right. The words and the speed were wrong — and those are the two things a system fixes better than a person does at 9pm on a Sunday.
What We Installed
The Referral Agent does the job Matt was doing manually, with two changes that the manual version could not hold.
The first is speed. The agent watches the groups and the coordinator list and surfaces a post the moment it goes up, inside the window where a reply still earns a read, not after it has scrolled past.
The second is the writing. Every first message references something specific to that coordinator's post or profile, acknowledges the difficulty in the placement they are describing, states one real credential, and makes a contingent offer. No "we'd love to help." The structure mirrors the cold email sequence the same engine runs: their situation first, the work acknowledged, the offer last.
We also moved the target. The community-access funding cut had reshaped the income model for community-only providers, so we leaned the messaging into SIL — 24-hour residential placements, which come almost entirely through support coordinators. Same coordinators. Same database. Higher-value placements. The funding cut made Matt's coordinator relationships more valuable, not less.
The First 9 Days
Nine days after go-live, five support coordinators had responded and raised their hands — opened a conversation, asked about capacity, or asked for the referral process. Five. From a provider that nine days earlier had a referral history of nothing.
— Build n Bloom — on what changedThe system did not make Matt more persuasive. It made him faster than the providers who reply slowly, and more specific than the providers who reply with a template. That is the whole edge. Coordinators do not hand-raise for the best pitch. They hand-raise for the message that reads like it was written by someone who saw their exact post.
A hand-raise is not a signed placement. We are careful with that word on purpose. It is a support coordinator — the person who controls SIL referrals — choosing to start a conversation with a provider they had never heard of nine days earlier. For a startup, that is the entire bottleneck broken. The placements come from conversations. Matt now has five, and the agent keeps the next ones coming while he runs the home.
The Numbers
| Metric | Before | After (Day 9) | |--------|--------|----------------| | Coordinator hand-raises | 0 | 5 | | Manual replies sent | 15 (0 converted) | Agent-handled, in-window | | Referral history | None | 5 live coordinator conversations | | Reply timing | When Matt was free | Inside the relevance window |
What this case proves is narrow and we will keep it narrow: a startup SIL provider with no history went from zero to five coordinator conversations in nine days, because the outreach got faster and the writing got specific. The agent did the part that does not scale by hand. Matt does the part that should never be automated — the conversation that turns a hand-raise into a placement.
If you are a provider whose referral pipeline depends on one person finding time to chase coordinators, this is the job the agent is built to take off your desk.