What we’ve learnt from deploying a chatbot for NHS 24 in response to Covid-19
Aly McKenzie
Lead Consultant - Business Strategist
BSA Competency Lead
Aly McKenzie, Ensono Digital Senior Consultant, shares how we provided a chatbot prototype in record time and our tips for anyone else thinking about deploying this technology.
As Scotland’s national provider of telephone and digital-based health services, NHS 24 has been at the forefront of providing information and care for people worried about coronavirus. The organisation has responded to the crisis by providing people with a number of options about the way they access information and advice.
One of those options is a chatbot – called Ave – that is able to answer questions about COVID-19 from the clinically approved information on NHS inform. At Ensono Digital, we engineered Ask Ave using the Microsoft Bot Framework. We collaborated with a number of other organisations, including the team at Cancer Central who came up with the the concept of a Q & A chatbot for NHS 24.
When we started working with NHS 24 in March, enquires from people with questions about Covid-19 were already high and set to grow. The pressure was on to develop a solution in record time. Ask Ave needed to help the organisation manage spikes in enquiries and provide accurate, reliable information that reflected frequently updated medical guidelines, while simultaneously enhancing customer experience.
We turned around a prototype in less than 24 hours and Ask Ave is now live.
So, what have we at Ensono Digital learned from our experience of deploying a chatbot under these circumstances?
Here are the five tips I’d pass on to anyone wondering whether their organisation could benefit from deploying this self-serve technology:
- Check your use cases. Do you need a purely informational chatbot? Or a chatbot that serves as a conversational UI? Accurately separate, categorise and prioritise your use cases so that you deal with the most valuable first, and work iteratively from there.
- Anticipate the spike. As with most things, prevention is better than cure. While it’s possible to deploy a chatbot during a spike, it’s much more comfortable to deploy one ahead of time.
- Adopt a content strategy that is chatbot-friendly. There are many ways to make it easier to train and update a chatbot to respond using content that your organisation may already have. Putting some effort into ensuring that such content is optimally formatted for use by a chatbot is time well spent.
- Get people involved. Content managers/creators, UXers, first line responders – these are the experts in your organisation’s message and image, as well as your audience and their needs. Get them involved and look for opportunities to upskill people to become chatbot champions.
- Know your customers. OK, KYC may sound obvious, but it’s an area which suffers a lot from assumptions being made. Be data-driven. Analyse how people interact with your traditional channels to build your understanding of how they will interact with your automated ones.
We’re used to partnering with organisations to help them plan and deploy the right technologies to enable them to quickly scale and meet changing demands. Right now, we’re supporting clients to deal with new challenges during the pandemic, as well as accelerating longer term growth objectives as part of their emergence strategy. If that sounds like you, please leave us a message below. We’ll be happy to share more advice and talk about how we can help your business succeed.
Social Share
Don't miss the latest from Ensono
Keep up with Ensono
Innovation never stops, and we support you at every stage. From infrastructure-as-a-service advances to upcoming webinars, explore our news here.
Blog Post | November 14, 2024 | Best practices
Leveraging Observability in SREaaS: Building a Robust MELT Stack
Blog Post | November 4, 2024 | Best practices
Navigating the Matrix: Proactive Mainframe Management Strategies
Blog Post | October 30, 2024 | Industry trends