Increasing numbers of public sector organisations are now doing rather than talking about bots, and they are beginning to see that words are important element of how they work.
Deployments of robotic process automation (RPA) and chatbots are on the rise, and there is a growing body of evidence around what does and does not work, the various elements of good practice and the priorities in a strategic approach.
A tipping point has been reached in which bots are moving into the mainstream of operational activity, ready to comprise a significant element of how organisations plan for the future.
This provided the main focus of UKAuthority’s recent Bots4Good conference, in which the emphasis was on the experience of some of the pioneer organisation and the detail of how to make bots work.
Amid a number of factors, the significance of words became apparent. RPA and chatbot implementations are designed to take on the more basic functions currently handled by human agents, language is an important element of the processes, and it is not an easy factor to incorporate.
Interpretations
A stand-out lesson from the presentations and discussions was that authorities are having to put a lot of thought into how their systems interpret words and phrases to ensure the appropriate steps are taken or responses given to statements. They have to acknowledge that the public is often unaware of its own terminologies and talks a different language in making requests.
They have to ensure that the common choices of words and turns of phrase are programmed into the systems, overcoming ambiguities and leaving room for minor errors. One of the examples highlighted during the conference came from a council that refers internally to ‘walking aids’ but had to recognise that the public is more likely to talk about ‘Zimmer frames’.
It creates a stiff challenge that takes time to overcome – with some experimentation and iteration along the way – but it is an important element of ensuring that bots operate reliably.
This is a factor that is likely to become increasingly important as automation becomes more sophisticated. As organisations experiment with machine learning they could need to spend even more time on those intricacies of language. But the conference conveyed the view that the long term benefits will make it a challenge worth addressing.
The full briefing paper from the UKAuthority Bots4Good event is available to read now. You can download the document from here. ADD LINK
Image from iStock, Abidal
