In February, we launched Multiverse Atlas - an AI tool that enables on-demand coaching, whether delivered by artificial intelligence or human experts.
We launched this tool to be there for apprentices whenever they have questions, with no delays: real-time, personalized, expert support delivered on-demand.
Atlas is our biggest step towards AI-enhanced learning, and the great news is that learners are using Atlas even more than we anticipated, and overwhelmingly finding it a helpful tool to understand course material, overcome challenges, and brainstorm ideas more efficiently. We're also building evidence that these technologies can be particularly useful for meeting the needs of some historically underserved communities, supporting more equitable access to high-quality training and education.
Atlas enabling scale
In a matter of months, 3,600 learners - that’s more than a quarter of our active learners - have asked more than 40,000 questions to Multiverse Atlas. With our bold ambitions to reach 100,000 learners before the end of the decade, it’s hugely important that we have confidence we can deliver personalized learning, at scale, and Atlas is enabling this.
While ensuring adoption was crucial, we wanted to make sure Multiverse Atlas provides genuine value and facilitates better learning outcomes, rather than driving superficial usage.
One of the measures we use to track this is a ‘response helpfulness score’ where we ask all users to say whether they find the responses provided by Atlas useful. The data showed consistently high helpfulness scores, slightly above 91%.
About half of the learners that message Atlas once will become frequent users: they are coming back again and again after seeing the value this on-demand coaching can offer.
Driving real business impact
While hard data is always going to be our guiding star, it’s qualitative feedback that really brings the impact of tools like Atlas to life.
Some highlighted the speed and accessibility as a big advantage. For example, one apprentice said they used Atlas to catch up after a hospital appointment which meant they had missed some sessions. Another said it helped keep them focused, as they could quickly locate answers without navigating through lots of different documents or waiting for a coach to reply.
Others emphasized the benefits of a flexible learning style. One person said they already liked to teach themselves by researching online, so this approach suited them very well. Atlas also allows learners to specify and personalize the type of responses they prefer. A number of users said they found this helpful as it felt harder to do with human coaches.
Ryan, a production controller at an aerospace company and an apprentice on our Data Literacy program, uses Atlas every day to solve functions and macros as he’s working with data. He said: “Whenever I’m stuck on something I can just pop it into Atlas and it works - it seems to be really good at interpreting what I need.
“I used to spend hours on YouTube following it step-by-step and I’d keep rewinding it back to the right bit so I can follow along. It was a really slow process and didn’t always work even then.”
These examples clearly demonstrate how Atlas is fulfilling our vision of blended, guided support that is delivering real impact for our learners and their employers.
Supporting apprentices at every age and every stage of their career
More than 50% of our learners are over the age of 30 on upskilling and reskilling programmes, and we know the benefit apprenticeships can have for those of any age. Any tool that we build cannot simply be there for so-called ‘digital native’ generations, it needs to be accessible and useful to everyone.
Contrary to assumptions that AI is grasped more intuitively by younger digital natives, Atlas has seen the highest adoption among apprentices over 40 years old. More than four in ten (46%) of learners in this age group have used it so far, compared to 31% of users aged 24 and under.
We think this is happening for two reasons. Our main hypothesis is that our younger apprentices are regularly using other AI tools (such as ChatGPT) already, as lots of research has found higher adoption among Gen Z and Millenials(opens new window). For many of our apprentices over 40, Atlas may therefore be helping to familiarize them with gen-AI tools.
During user research interviews several older apprentices also said they liked the anonymity it provides and allowed them to ask questions they might have been embarrassed to ask otherwise. This is a good example of where Atlas can help remove perceived stigma around knowledge gaps.
The equity imperative
Research(opens new window) and media coverage(opens new window) scrutinizing the risks of AI bias, continually shows how these systems can accidentally reproduce and amplify societal prejudices and inequalities from the real world, while accessibility is often an after-thought. From the outset, we were determined Atlas would be different.
Atlas was designed with a big focus on accessibility - including by following AA level Web Content Accessibility Guidelines - to make it easy to use for those with a variety of disabilities. This included features like keyboard navigation and screen reader compatibility. Strong guardrails were also built into Atlas to reduce biases and ensure professional and contextually appropriate responses.
Atlas has seen slightly higher adoption rates among those with additional learning needs. 46% of learners with critical additional learning needs have adopted Atlas so far, compared to 37% with no additional needs. This clearly demonstrates the value of investing in accessibility.
We're also seeing equitable usage across ethnicities, with similar adoption levels for Asian (37%), Multi-racial (36%), and White (41%) apprentices. However, Black apprentice usage is lower at 32%. We believe this can be attributed to the higher Atlas take-up among our older apprenticeship cohorts, which are less ethnically diverse than younger groups. However, we’ll be investigating further to make sure there’s nothing else that could be contributing to the difference.
Gender adoption has been pretty much equal too, with 38% of male apprentices and 40% of female apprentices using Atlas to date. This is significant as a study carried about by the University of Chicago(opens new window) found that “women are about 20 percentage points less likely to use ChatGPT than men in the same occupation”.
A continuously improving, context-aware companion
While we’re encouraged by these initial results, we're just beginning our exploration of Atlas's potential as an AI-powered learning companion.
Our long-term vision is for Atlas to become an advanced, deeply personalized, and context-aware AI coach. Working alongside human coaches it should be capable of supporting apprentices at every step of their journey towards competency, mastery, and career success. We've already making progress and identified where we’ll focus next:
- Continue to closely track user experience data and gather in-depth feedback from apprentices and coaches. Their perspectives matter most in shaping Atlas's future development.
- Make it easier for apprentices to personalize and fine-tune Atlas to their own preferred learning styles through tailored settings and configurations.
- Enhance Atlas with advanced capabilities to provide hyper-contextual support mapped to each learner’s precise journey and progress through their apprenticeship program.
- Integrate Atlas more closely within the workflows of our human coaches teams with auto-suggested tailored content and guidance to facilitate more seamless hybrid human-AI instruction.
We're excited to build a future where everyone can access the personalized learning support they need to enable support upskilling and career growth - regardless of age, ethnicity, gender, wealth, or learning style.
Through AI powered tools like Atlas, we’re also able to do this at even greater scale, helping lots more companies to deliver continuous learning and close their skills gaps as they prepare for the AI-enabled future.
And just as we ask our apprentices to commit to their learning journeys, we'll continue our unwavering commitment to place equity at the heart of our approach to AI in education. We do this because it’s the right thing to do, but also because it’s needed if we want to close the large skill gaps that exist right across society.
