One school sends school-wide information and timetables through a school information portal. The portal's app notifications are broken. And when teachers use it, they don't always send the information in the announcement itself. It's in an attached PDF, so if you're not reading all of it, you're missing something.

That school also puts all the homework and classroom discussions on Google Classroom, and not every teacher uses Google Classroom in the same way.

The other school has a different timetable, a different lunch menu, and its own WhatsApp groups and Telegram channels.

Schools send information to us in eight places: the school information portal, Google Classroom, three Telegram channels, two WhatsApp groups and a lunch-menu website. That's a lot of channels for parents to coordinate and understand.

Who needs this information

As parents, we need to know what each school has sent. One of the kids needs help navigating their homework and making sure it's done on time and to a high standard.

Breaking it down

For each of those eight sources, I have to work through the same six questions:

  1. What information is available?
  2. What is the current format that it is in?
  3. What format can we read it in? What format do we ultimately want it in?
  4. Where can it be stored securely and safely? (Also, are we only taking the information that we need?)
  5. How can the information be transformed?
  6. How can the information be queried and interacted with?

What I want it to do

Collect all of it on a schedule. Put it somewhere it can be queried. Send new information to us on Telegram. Let the Telegram bot answer questions about that data, and update it, for example from a photo of a new timetable.

A process flow diagram running left to right: eight labelled sources (portal, classroom, three channels, two groups and a lunch menu) are collected on a schedule into one store on Cloudflare OS, and one Telegram bot pushes new information to parents and answers their questions, with a loop back for parents to update the store from a photo of a new timetable.

Then you have a better method of working with the multi-platform data, but from just one point of interaction for the people who need to read it.

The first version went live earlier this month on Cloudflare OS. Next week I'll answer the first two questions for each of those eight sources: what information is available, and what the current format is.

CloudflareAgentsKnowledge managementProductivity