17th Oct 2022

Customer Relationship Management (CRM)

We have recently been trying to integrate a new Customer Relationship Management (CRM) system into our workflows. The goal of this system is for us to manage our communication with customers of our various services in one place.

Ryan and I spent more time this week problem solving some issues with the CRM. We ran into quite a few road blocks with the CRM but we managed to have a call with the our CRM provider who connected us to the solutions. Still plenty work left to do but we are still making good progress.

TTS Systems

Text to speech (TTS) is what is used to generate synthesised voice by a communication aid or any computer.

We have been hearing a lot of issues come up recently from our own clinicians and from external teams to do with TTS systems. We regularly hear about a lack of regional voices, or regional voices exist but are locked to a system that won't work for a client. We also hear the same thing about voices for various languages, they exist but are not in the system we want to use. This problem is incredibly frustrating as we can see the solution, we just can't use it.

We also know there are lots of ways to generate custom voices based on voice recordings. However, currently it is expensive to create these and then the voices is usually locked into a given system. Given that the technology exists to create these voices we want users to be able to take their voice to whatever system works best for them, even if that changes.

From a developer perspective it can be difficult to work with TTS systems, which in turn causes the above issues. If you are a small team making an new AAC app, for example PASCO. You have to choose which TTS Voices you make your application work with, the more voices you want to offer the more work you have to do. Some of this work might have to be redone if you want your app to work on different platforms. When teams are faced with this choice they usually have to settle for a small selection of voices. This is what leads to applications only supporting a subset of voices. All these small teams are also doing the exact same work to integrate with TTS Systems.

So with these problems in mind we want to see how we can work with the development community to break this cycle and get everyone working together on having voices across platforms. In an ideal world we would create an Open Source TTS System which integrated with all the existing clever technology. That way the developers could integrate with one simple TTS System but unlock every voice ever made for their users. Wouldn't that be cool?

Anyway, this is a pretty big challenge and we really have no idea how we will go about this yet. So we are still exploring the options. This week I have really just been getting to grips with the technology behind TTS systems so I can actually understand the problem more deeply and propose sensible solutions. From what I have learned this week, many people have tried to tackle this problem and its hard.