Four years ago, I wrote a blog focused on the emerging world of artificial intelligence (AI) coaches, exploring the massive enhancement they could bring to learning experiences in terms of practice and reinforcement.
At that time, the AI coaching landscape was in its early days; no solutions had gotten massive traction yet—either on the consumer or enterprise side.
The most effective solutions were in the public speaking niche. Apps such as Orai, Speeko, and Virtual Speech allowed people to practice speeches or presentations and get instant statistics and feedback. These could be used alone (to democratize basic access to coaching) or in conjunction with professional human coaching to extend and practice learnings. Solutions focused on more generic leadership were few, with fairly clunky user experiences. This was an interesting trend but not a massive disruptor quite yet.
Fast forward to the present. We now are seeing a new wave of AI coaches, powered by the explosion of Large Language Models (LLMs), which are much more poised to transform learning solutions. ChatGPT has been a pioneer in this space for consumers, able to play the role of any type of coach—from sales to fitness, healthcare, leadership, and life in general. Unlike traditional bots, Apple also is adding AI capabilities to Siri to make it more personal and context-specific (it was still unreleased at the time of writing).
ALWAYS-ON TRAINING PARTNER
Amid this landscape, companies are re-exploring AI coaches as part of a broader, holistic learning experience. These bots allow for chat- or voice-based practice and application—like an always-on mentor and training partner.
A popular use case, for instance, is sales and customer service training. Bots can be trained on company frameworks and tool sets, as well as asked to play the role of various customer personas. Learners then can practice navigating challenging customer and sales situations through natural conversation (vs. choosing pre-made options, as in traditional scenario-based learning). The bots then can flip from the role of a customer to a coach, providing customized feedback and tips for improvement.
The look, or embodiment, of these bots plays a crucial role in shaping the user experience. While bots such as GPT voice or Siri adopt an abstract, waveform appearance, scenarios requiring a heightened sense of empathy—such as customer interactions—often leverage more human-like avatars. Providers in this space include Soul Machines, DeepBrain AI, Athena from ENGAGE, MetaHuman from Unreal Engine, and Liteforms from Looking Glass Factory.
An insight from working with clients is that despite the common desire for an ultra-realistic avatar, it is often useful to start with a more stylized (“character- like”) one to avoid the uncanny valley effect—a feeling of unease due to subtle differences with humans. This effect can pop up if the bot looks realistic but is slow to respond or if its lips are not fully synced with speech due to bandwidth or API lags. A less lifelike avatar, on the other hand, is more forgiving with subtle imperfections. These are the kinds of quirks and nuances to discover through critical hands-on exploration.
A SUBSTANTIAL LEAP FORWARD
In 2024 and beyond, AI coaches are poised to evolve into smarter, faster, and more integral components of learning programs and coaching experiences.
To stay at the forefront, Learning leaders must proactively explore ways to integrate these solutions into their workflows. An accessible starting point is the ChatGPT app, especially with its voice capabilities, followed by avatar bots from Liteforms or Soul Machines. These offer customizable data and personality training, along with trial versions of their products, providing an excellent avenue for testing and learning.