Skip to content
All posts
Product

What a good follow-up call actually sounds like

Marcus ElleryCo-founder & CTOJune 3, 20265 min

There is a version of an automated call that everyone has suffered through. It talks over you. It reads a script. It asks you to press one. It does not listen, and you can feel that it does not, so you hang up and nothing was learned.

A follow-up call after a hospital stay cannot be that. The person on the other end is tired, maybe in pain, maybe a little scared. The call has to earn the next thirty seconds, over and over, or it is worse than no call at all.

It says who it is, and why it called

Tend opens by naming itself plainly: an assistant calling on behalf of the patient's care team, to check on their recovery. No pretending to be a person. People trust what is honest with them, and honesty is also the safest design choice.

It moves at the patient's pace

  • It waits. Silence is not a failure state; it is someone thinking.
  • It lets people finish, and it can be interrupted without getting confused.
  • It repeats anything, as many times as needed, without a sigh.
  • It can hand off to a person the moment someone asks.

It knows exactly what it is not

The most important sentence in the call is the one where it declines to give medical advice, and offers to bring in someone who can.

When a patient describes something that could be a complication, the call does not reassure and it does not diagnose. It acknowledges, it records carefully, and it escalates to the care team. The warmth is real, but it never blurs the line between checking in and practicing medicine.

Get those small choices right and something surprising happens. People talk to it. They mention the thing they were not sure was worth mentioning. And that is the whole point of the call.

Tend is voice AI for post-discharge follow-up. It calls patients after they leave the hospital, checks in warmly, and flags concerns to the care team early. It does not give medical advice. It checks and it escalates.