Launching on-demand coaching, powered by AI

Artificial Intelligence is creating new ways to add value to our learners and customers, and better routes through which to do it.

First, we’re equipping people with the skills they need to thrive in an AI age. We’ve already trained thousands of individuals in advanced AI skills via our data programmes, and last summer we rolled out AI training to every one of our apprentices.

And secondly, we see AI as a solution to some of the biggest challenges that education has faced to date. World-class training has been held back by scarcity: but the strength of AI is its ability to turn scarcity into abundance.

It’s why we’re investing heavily in our technology, to develop the tools that will unlock outstanding training to hundreds of thousands of learners.

This month, we’re launching one such tool: on-demand coaching, powered by AI.

Expert, human coaches have always supported our apprentices, and will continue to do so. But we wanted to create something that could build on that support. To be there for apprentices whenever they have questions, with no delays. We know that, for example, our apprentices have a lot of questions as they approach crunch moments, projects and assessments - and those questions don’t always come during working hours.

So, during one of our regular hackathons, our tech team built a first iteration of an always-on, AI-powered assistant coach. It was immediately clear that this could benefit apprentices as a first port of call: to help them understand course material, overcome challenges, or brainstorm ideas.

Cem Gurkan, Product Manager, said: “As the world of work becomes more flexible and we push for more inclusive workplaces: not every assignment is done from 9-5 any more. The rapid acceleration in AI technology can make it possible for us to be there for apprentices whenever they need us.”

Enter: Multiverse Atlas.

Multiverse Atlas is built on a commercially available Large Language Model (LLM), and has been carefully prompted to support our apprentices.

It’s designed to encourage a socratic method of coaching, aiding apprentices to delve into topics themselves and find solutions, rather than simply giving answers like an off-the-shelf chatbot might. Atlas can give career advice, answer questions on topics related to our programmes, and quiz apprentices on topics - helping them to study.

Clare Dodd, VP Global Delivery, said: “Our coaches are industry experts that go through a rigorous coaching academy that covers our coaching philosophy: how we teach apprentices to ensure they are empowered to do their best work, retaining and applying their skills and knowledge. Atlas enables coaches to do what they do best, whilst also providing assistance in the moment an apprentice needs it most.

“For Atlas to do that successfully, the learning team worked closely with the tech team to test and iterate Atlas’ style of responding to apprentices - guiding them to find answers, and referring them to their human coach when they need additional support.”

The prompt that powers Atlas went through more than 100 iterations over a three month building process, and it continues to evolve with the close input from our coaches.

Atlas tailors its responses to the apprentice it is speaking to: it knows who they are, the industry they work in, the sort of job they do and the apprenticeship programme they are studying. At the core of what we do is the belief that people learn better when content is personalised to them, and AI enables that at scale.

In our beta tests, more than 10,000 questions have been asked and apprentices found 90% of Atlas’ responses to be helpful. For when Atlas doesn’t have the answer, apprentices can speak directly to their coach from the same on-demand chat application. Coaches can also see Atlas’ responses and can clarify or correct. On-demand coaching ultimately enables apprentices to connect directly with support, no matter where they are.

This is just the beginning. Multiverse Atlas, and our on-demand coaching, will keep evolving.

As you read this, Atlas is being trained on all of our courses: so it will soon know our programmes like Software Engineering or the Data Fellowship back to front, to guide apprentices through their work. We’re continuing to personalise Atlas to each learner, so it can provide advanced contextual support based on where they are in their programme. We want to build the ultimate assistant coach that journeys with apprentices and provides them helpful, tailored assistance as they learn.

Peppa Wise, VP GTM, said: “When we’ve asked business leaders, they’ve told us that they don’t have the skills they need to reap the productivity gains AI could bring. Through our module - AI Jumpstart - we're offering training to all of our apprentices in these technologies, to start to close that skills gap. And AI tools themselves will enable us to make that training even better, and reach even more people: so more businesses can see the benefits that AI will bring.”

And Atlas is just one part of how we’re thinking about AI. Looking to the future, we’ll launch more ways to ensure our apprentices are power-users of AI tools, and experts in machine learning: while building the tools that will enable world-class, personalised, applied learning at scale.