.jpeg)
A member finishes signup. Your scheduler has no first appointment for them. Before you send another reminder, you need to know whether they reached booking at all.
Hello Patient builds AI agents for healthcare. Our agents can call or text within agreed outreach and consent rules, answer administrative questions from approved information, and send personalized booking links. Deciding where that help belongs starts with a measurement question: where, exactly, did members stop?
For a virtual mental-health platform, signup, matching, booking, and attendance need separate counts. This is one part of the broader member journey.
Where do members stop?
Follow unique people through one selected path. Choose an entry-date range and give each person the same follow-up duration, measured from their original entry event. Count later events only within that person's window.
This recommended worksheet follows a path that requires matching. Use a separate worksheet for direct booking, with matching omitted. Before counting, define the local events: a qualifying inquiry, completed signup, readiness for matching, completed match, readiness to book, first booking, and recorded attendance. In this example, a completed match also establishes readiness to book. If your process separates them, add that handoff as another row.
Every row below includes only people in the original inquiry cohort who reached that row's entry event. The owners are suggestions to adapt.
| Stage | People entering | People advancing | Suggested owner |
|---|---|---|---|
| Inquiry | People with a qualifying inquiry | Those completing signup | Growth or inquiry team |
| Signup | People with completed signup | Those ready for matching | Member experience |
| Match | People ready for matching | Those with a completed match | Matching team |
| First booking | People with a completed match, ready to book | Those with a first appointment recorded | Scheduling |
| First attendance | People with a first booking | Those with a first attended session recorded | Care operations |
For each row, divide people advancing by people entering. If nobody entered, show “not calculable.” These are person-level conversion rates. Their complements are not appointment no-show rates, which need their own appointment denominator.
Keep appointment outcomes distinct: recorded attended, confirmed missed, canceled, rescheduled, future, and unknown attendance. Retain the history when someone reschedules and later attends. Unknown attendance means the outcome is unconfirmed; an unknown reason means you do not know why someone stopped.
Also report the share of the original cohort reaching a first attended session. Keep people with unresolved outcomes in that original total and disclose their count. A reply, sent link, or click cannot establish booking or attendance. Use scheduling and attendance records.
Which barrier is responsible?
Use the stage with the loss to narrow your investigation. The count tells you where to look; it does not explain why someone stopped.
For members ready to book, review available appointments alongside the booking flow and any questions they raised. Provider availability, unsuitable times, unclear costs, a confusing flow, and personal readiness are hypotheses to investigate.
Scheduling or care operations can investigate access. Product can investigate repeated booking errors. Uncertainty about starting care may call for a human conversation about what the member needs. Adapt those assignments to your team.
If members receive a link but no booking follows, check what it offered at the time. Could they find an appropriate provider and an appointment they could attend? Did a question need answering first? Record an unestablished reason as unknown. Never interpret silence as a refusal of care.
If the loss occurs during intake, examine those mechanics separately: what an AI patient intake agent does and how it differs from a portal form.
What can administrative follow-up handle?
Administrative follow-up can help a member get an approved answer or reach the next scheduling step.
Within agreed consent and outreach rules, our agents call or text members, answer approved administrative questions, send personalized booking links, and create tickets for your team with the member's details.
Our agents can state accepted plans from approved information. Clinical questions go to the care team under the platform's protocol; agents do not give medical advice.
Set the outreach scope, consent rules, and escalation path before testing. Our agents honor opt-outs right away.
Hear how it sounds, or request a free demo agent for an administrative conversation you want to examine.
What does the evidence show?
In its published case study, Octave reported nearly doubling its capacity to answer inquiry calls four months after rollout. Read the Octave case study.
A 2024 retrospective study at the University of South Florida compared telemedicine and in-person appointments from January 2021 through April 2023. In the matched cohort of 87,376 appointments, completion was 73.4% for telemedicine and 64.2% for in-person appointments. This was a health-system comparison, not a study of first sessions at a direct-to-consumer mental-health platform or an AI follow-up intervention. It does not provide a universal normal no-show rate for your members. Read the study.
How can we test a change?
Test one change at the stage you identified. Write down the eligible group, proposed change, outcome, and observation window before the test begins.
Where feasible, use a comparison group that retains standard service. Allow enough time to book and attend, and compare equivalent windows. A before-and-after improvement alone cannot establish causation.
Show first bookings and first attended sessions alongside opt-outs, escalations, and cost. Include provider availability and clinical capacity limits in the result. More clicks with unchanged attendance should prompt another look at the path.
Schedule a call to discuss one member path. We can use your stage counts to locate the loss and decide where agent follow-up belongs.
Frequently asked questions
How should a virtual-care platform count a first session canceled by the provider?
Keep provider cancellations separate from member no-shows when evaluating follow-up by Hello Patient. Have your platform team record who canceled, whether another appointment was offered, and whether the member eventually attended within their observation window. The agent can send a personalized booking link or create a ticket for your scheduling team with the member's details; your records establish the outcome. Keep the member in the original cohort, including when the outcome remains unresolved. Recording cancellations separately helps your team identify a capacity or scheduling problem that may need a different response from member outreach.
How should we handle duplicate virtual-care signups when measuring first-session attendance?
Your platform team should reconcile duplicate signups before measuring first-session attendance after Hello Patient follow-up. Use your approved identity process, retain the underlying records, and count each person once in the selected cohort. Decide how repeat signup attempts relate to the original entry date and apply that rule consistently. Do not assume two similar names belong to the same person. Flag uncertain records and show how they affect the totals. The agent can answer approved administrative questions and create tickets for your team with the member's details.
Does a rescheduled first virtual-care visit count as a failed first session?
A rescheduled visit needs its own appointment status in your platform's records when reviewing Hello Patient follow-up. Measure whether the member reaches a first attended session within their observation window, and retain canceled, missed, or rescheduled appointments separately. Someone may reschedule and then attend; counting only the original appointment would miss that progress. Moving an appointment does not establish attendance, either. If the new date falls beyond the cutoff, show the future appointment and the unresolved attendance outcome. Use the same follow-up duration for each member in each comparison group.
How long should we observe virtual-care members before measuring first-session attendance?
Use your scheduling lead times to set the window for reviewing Hello Patient follow-up. Have your team inspect how long members currently take to reach their first booking and first attended session, then select a period that allows those outcomes to occur. Report entry dates and the follow-up duration together, measured from each person's original entry event. Identify members who still have future first appointments. Avoid comparing a cohort with complete follow-up against newer members who have had less time to attend.
What if our virtual-care signup volume is too low to compare first-session attendance?
Start with individual member paths when evaluating a small test of Hello Patient follow-up. Have your platform team identify each eligible member's last recorded stage, any known barrier, the action taken, and the eventual scheduling outcome. A handful of cases cannot prove an intervention works. Keep unknown reasons separate from unknown attendance, and avoid splitting a small cohort into many smaller segments. You can check whether agents send the intended links and create tickets for the right teams while collecting more observations, then measure attendance from your platform's records.
What if a virtual-care member asks to change channels or avoid private details on a shared phone?
Agree on a route for contact preferences before Hello Patient outreach begins. Decide what happens when a member asks to switch channels or says they share a phone and want to avoid detail there. The agent follows the route you set, stays within the consent scope you agreed, and honors an opt-out right away. Do not assume a reply permits further detail on that channel. For more context, read privacy and security guidance.
.svg)
