Behind each reply
LiveBehind each reply.
Running since September 2026 · Updated November 10, 2026
Every figure below comes from Phia’s own records: when each text arrived and was answered, and what Phia checked, recorded, planned and sent each time it ran. Only percentages and typical waits are shown, and the page updates once a day.
10 sec
typical wait between her text and Phia’s reply
90%
of her texts got a reply within a minute, from Phia or her care team
15 sec
typical wait when she texted at night or on a weekend
Chapter 1
What happens when she texts
Every time she texts, a follow-up comes due, or something changes in her chart, Phia wakes up and works through the same steps. The reply she sees is the last of them.
01 · wake
A text arrives
A new symptom, at 9:40 at night
02 · check
Danger signs
Any warning sign goes to the care team first
03 · record
What changed
The symptom goes on her timeline
04 · plan
The next touch
A check-in for tomorrow, decided again then
05 · reply
Then it replies
The text she sees comes last
Chapter 2
Most of the work isn’t the reply
A reply is the part she sees. Around it, Phia keeps her record current, checks what needs checking, and lines up whatever comes next, so the next text starts from where she is.
A run is each time Phia wakes up and decides what to do. Many runs are a simple answer to a simple question, so each kind of action shows up in some runs rather than all of them.
Figure 2.1
What Phia’s runs included
Share of runs that included each kind of action. A run can include several, so shares don’t add up to 100.
Keeping track
Recorded a milestone or symptom, or planned when to check in next
22%
Hand-offs
Passed the conversation to a person on her care team, or looped in a specialist
11%
Safety
Checked for danger signs, ran a short safety screen, or raised an alert for her care team
7%
Paperwork
Sent intake, consent or a screening, or checked her service area
7%
Logistics
Appointments, phone calls, and help with food, housing or transport
7%
Guidance
Looked up clinical guidance or her care plan before answering
under 5%
Chapter 3
What it was for
Each text she sends is also sorted by topic, separately from the reply. Here is what the texts that set Phia in motion were about, and how it followed through.
Figure 3.1
What her text was about
Share of runs started by a text from her.
41%
When she texted about a symptom, Phia added it to her symptom record for the care team to track.
31%
When she texted about a symptom, Phia checked for danger signs or raised an alert for her care team.
28%
When she texted about logistics, Phia handed the conversation to a person on her care team.

Chapter 4
Knowing when not to text
When Phia reaches out on its own, its first decision is whether she needs a message at all. A check-in she doesn’t need is one more thing on her phone, so often the right call is to wait.
Figure 4.1
How Phia’s own check-ins ended
Share of runs Phia started on its own, by whether it texted her or decided to wait.
Scheduled check-ins
Chapter 5
Working alongside the care team
Phia doesn’t work alone. The care team writes in the same thread, and Phia brings them in when a conversation needs a clinician.
88%
of the time, the first reply to her text came from Phia
2 min
typical wait when someone on her care team replied first
Figure 5.1
Who wrote the texts she received
Share of texts sent to her.
19%
of women who text Phia have been looped in with a specialist from her care team
35 min
typical time for a clinician to pick up a loop-in
Figure 5.2
Who Phia brought in
Share of loop-ins by the kind of specialist she was connected with.
Mental health
70%
About these numbers
- Drawn from Phia’s own action log, with test and staff accounts removed.
- Topics come from what she wrote; actions come from what Phia did next.
- A wait runs from the first text she sends to the first reply that reaches her. Texts sent back to back count once, and a scheduled check-in is not counted as a reply. A typical wait is the median.
- Only percentages and typical waits are published, never totals. Every figure stands for at least 11 women, and no text, name or reasoning is ever shown.
- The page updates once a day.
















