About SchoolBot

I am NOT getting involved in school communications…but…

I have school children now. And I kinda made a vow to not get involved in school communications. Mostly because the primary job of school staff is educating and caring for children to do, and they don’t want some geeky parent turning up with bright ideas that will probably make their lives more complicated. In short: I don’t want to cause trouble.

But like churches and councils and various other public organisations that do way more than just sell stuff, schools have difficult communications challenges. They have lots of information about lots of different things, and any one person is only interested in a small amount of it.

Organising and presenting this information such that people can see what they need without the stuff that they don’t need getting in the way is very difficult. And, while I have some ideas for improving it, I don’t have ideas for fixing the whole problem.

I’m a parent now, and so getting the information that a parent needs for day-to-day schooling is what concerns me most.

One interesting way in which I see parents trying to get the information they need is to collaborate. Parents’ Facebook groups pop up, and parents ask each other questions in these groups, and with several hundred parents, answers can arrive quite quickly. This is great!

But I’m also a software developer, and so when I see the same problem appearing repeatedly and being solved repeatedly, my instinct is to make it more efficient.

So when I see questions that are frequently being asked, and people having to wait for answers, I wonder how we can get people the answers more efficiently.

Surveying the landscape

I’m not a school comms expert, so there’s probably lots missing here. But from what I can tell there are:

  • Tools that send information to parents, like Schoop and ParentMail
  • Tools that contain some information, but who’s primary purpose is elsewhere, like ParentPay
  • Generic content management systems, like WordPress or Drupal or paid services like SquareSpace that can be used for a website for any kind of organisation. These systems are not school-specific.

What I’ve been unable to find is a school-shaped content management system that stores and presents information that parents and other people interested in a school might want to see.

There are a few things that seem to advertise themselves as content management systems for schools, but they don’t seem to be huge advancement on standard CMS’s like WordPress and Joomla.

And anyway, I don’t think I wanted to a system for managing “content” as much as a systems for managing “information” – information that is specific to schools that can be presented in a variety of different ways, in order to fix the challenge of answering repeated similar questions.

Enter “SchoolBot”

I think the idea of SchoolBot came before my ideas about an information system emerged from it. I didn’t want to get involved in my own kid’s primary school website, but I wanted to make the information on it more easily accessible.

The people were asking questions on Facebook. And I’d been tinkering with Facebook Messenger “Bots” on another project. Bots are computer programs that can automatically respond to people using Facebook Messenger or other instant messaging systems.

So I wondered if I could create a Messenger “bot” that could use information from the school website to answer peoples’ questions without them even leaving Facebook – which for many people is the main way that they use the internet.

Initial schoolbot tests (on a non-school day!)

The initial SchoolBot was built in just a few hours, but Facebook Messenger was on a bit of a lockdown in the wake of the Cambridge Analytica investigations so it wasn’t available on Messenger to start with. And it was limited to a couple of initial functions like telling you the school meals for the day, and giving you term dates.

Over time, SchoolBot has become able to answer more questions, and generally be smarter. And it’s now connected to Facebook Messenger too.

If you’re thinking it, this is NOT “Artificial Intelligence”. It’s actually really quite dumb. But it’s clever enough to have the veil of something intelligent-looking.

And as a parent – and yes, I’m currently the only parent that has access to it – it’s a super quick way to ask those essential parent questions and get responses in super quick time. And it’s fun too!

A central “brain” for parent information

SchoolBot did start out as just a simple parent-information chatbot. But it had the information behind it stored in a small database.

I quickly realised that I could use this information in other ways that could be really helpful.

With the information in a database, I could not only respond to messenger questions, but I could generate text or tables for a website, or even – given a suitable template – printable PDFs!

A test of getting data from SchoolBot into a web page – this uses actual SchoolBot data

Both of these things are important. I totally understand the need for things like school meal menus to be in printable format for people that need them. And it seems that most documents start as PDFs, then these are uploaded to the website – which isn’t great because PDF’s on websites are really hard to use, especially on mobile devices.

So rather than having the information in MS Word and on the website and in the SchoolBot database too (and possibly other places), you could just have it in one place – the central SchoolBot “brain” – and all the other things can be generated from that!

Technical geekery note: yes, the aim here is to have the data available on an API and create widgets and so on that consume that API – in the screengrab above I’ve built a simple WordPress plugin that uses SchoolBot data in shortcodes. And yes, the holiday dates are derived from the term dates! You didn’t think I’d store the same information twice did you?!

SchoolBot also gathers feedback. It keeps a tally of how many times it has been asked questions that it can answer. And questions that it does not know how to answer are logged so the system can be improved by feeding it more data! Both of these things could be used by a school to improve parent communications by helping them recognise what parents need to know, or what information is hard to get hold of.

The impact on community

In all of this I’m really aware that, if this ever actually becomes a useful thing, I might be taking something important away. The Facebook group is a source of community.

People answering other people’s questions is a way to connect parents outside of the school run. Who knows what relationships might have sprung up just by someone spotting someone else at the school gate and saying “Oh, hi. You helped me out the other day!”

I don’t know if the benefits of quick answers to questions outweigh the potential negative effects of removing the Q&A from the group.

I don’t even know if the messenger bot part of SchoolBot will be something that people use. Though the centralised data aspects of it remain useful, even without the messenger-related parts.

It still feels like a useful thing. So I’m going to forge ahead and see what I can make of it.

What’s the status of SchoolBot?

It’s very much in development, but like many of my side projects, it exists in a somewhat-usable form and competes with other projects for my attention. Currently only I can use it. But there’s no reason why, in its current form, it couldn’t be put to the public to test.

But that would involve me being a geeky parent turning up with bright ideas that will probably make teacher’s lives more complicated.