Our transition to a distributed model completely changed the way we conduct training. One of the biggest challenges of going online-only was how we could replicate the interactive practice scenarios we offer our volunteers. Our work with Proxima previously was more than enough to convince us that they could provide a perfect solution to this problem, powered by conversational AI. What has resulted, is a new tool for us to use to train volunteers in a flexible, scalable and effective way. At a time when we are seeing a substantial increase in the demand for our mental health support services, our partnership with Proxima has been absolutely vital.
Lifeline is a non-profit organisation providing free, 24-hour mental health support services in Australia.
The 2018 evaluation of the Lifeline Text pilot demonstrated that a Lifeline text-based crisis support service can achieve positive outcomes for help seekers and increase accessibility of its services. As a result, in 2019 Lifeline began the extension of its Lifeline Text service.
However, a larger workforce of volunteers was required to operate this service. For the first time in Lifeline’s history they were required to recruit and train a 100% online and remote workforce. ‘The Lifeline Text: Training Program’ is required to efficiently and effectively induct, educate and assess a new remote volunteer workforce.
The team recognised that there was a pressing need to replicate and scale the important practice scenario training that had previously been carried out in face-to-face training. In October 2019, we were appointed to build a solution - a new simulated digital training tool that would adequately equip Lifeline’s volunteers to meet the ever-growing demand for its critical services.
Lifeline’s shift to a distributed volunteer training and service delivery model completely changed the way the organisation was required to train its planned 1,500 Text Crisis Supporters.
Whilst online training provided many of the skills and theory that volunteers needed, the critical face-to-face scenario training was a missing component. Replicating these practice scenarios was vital in order to fully prepare Lifeline’s volunteers to safely and effectively deliver critical suicide prevention crisis support.
Fast forward half a year and the need for training a distributed volunteer network had become even more pressing, as Covid-19 swept the country. Aside from the obvious physical impact on people, the pandemic has had an enormous impact on people’s mental wellbeing too.
Ongoing uncertainty, lack of relational contact and numerous other factors, have all contributed to a significant spike in mental health struggles across Australia and elsewhere. In response, Lifeline’s services have become more valuable than perhaps ever before.
The organisation’s volunteer Crisis Supporters provide suicide support and emotional assistance, whether over the phone, by text or online.
Delivering a usable, scalable, and effective method of practice scenario training, not only to transition to Lifeline’s new distributed model, but also to meet the exponentially growing demand for its services, was critical.
For Lifeline, this fundamental shift in its approach to volunteer training presented a unique challenge to think creatively about the way it did so. Building an automated digital training tool that would provide volunteers with vital practice and experience in a simulated crisis support scenario could make training more flexible, repeatable and scalable.
Having worked with Proxima before, Lifeline knew that our engineering team would be able to both conceptualise and deliver a solution like this, preparing its volunteers in a completely new way.
For Proxima, our new partnership with Lifeline provided an exciting opportunity to develop an entirely new concept and deliver something we’d never built before.
Developing a brand-new solution is not without its challenges.
The brief was to deliver a training simulation for volunteers that, powered by conversational AI, would closely replicate the types of conversations they would go on to have with real people, once trained.
Creating conversations that were lifelike, helpful, and sensitive to difficult topics was a key consideration.
Additionally, it was vital that we could offer clear channels for feedback to the volunteers being trained, allowing Lifeline to capture, review and analyse interactions.
A unique solution was needed.
Together with Lifeline, we got to work on building two interactive AI training conversations that simulated the real-life exchanges volunteers might encounter.
Embedded as a webchat into open-source learning management system Moodle, the two conversation profiles are designed to run anywhere between 8-15 minutes in length. They include multiple different branches in the conversation, in order to replicate a crisis seeker steering off-topic. Offering detailed human-like behaviours, such as pauses, typing indicators and redirection, the two conversation profiles are also language adjusted to reflect differences in user language, personality and background.
With the two conversations varied and repeatable, trainees are able to practice multiple times and build their confidence in a safe and flexible environment, before going live with real crisis support seekers. The Lifeline training team are also able to review all conversations in Moodle, providing volunteers with important assessment and feedback through the platform to help develop their skills and confidence.
This unique learning solution allows Lifeline to continue to scale their digital training offer, and run online training for any number of future volunteer cohorts.
With technology coming under ever-increasing ethical scrutiny, we are proud that we have always sought to develop solutions that have a genuinely positive impact in the world. Partnering with Lifeline to develop unique crisis support training tools, in a season unlike any other, is certainly one example of that. Excitingly, we believe that this is the first conversational AI tool to be deployed in mental health service training, and the opportunities for us to have a positive impact in this area are truly staggering.
With 3,000 volunteers to train each year, our solution has already become a valuable training tool in Lifeline’s kit. The solution has added a new level of accessibility and flexibility for training remote volunteers, helping the organisation to equip people across Australia to respond to the increasing demand for Lifeline’s services.
For Proxima, this was another successful evolution in our product line. Offering AI-powered training will be increasingly important for businesses of all shapes and sizes. If it’s something you think your team would benefit from, don’t hesitate to get in touch.