Using AI to overcome workplace biases

In the summer of 2022-23, founder Melissa Jenner sat down with a friend to discuss AI. 

At the time, she headed START Now, a human innovation company helping people and businesses create more purposeful, sustainable careers and work. She started the venture in 2015 after “rage quitting” a bad corporate job and realising the traditional methods of coaching were not only expensive but unreliable. 

“It’s hard to open up to people you don’t know and don’t trust. Everyone has their own frailties and biases.” 

As the friends waxed lyrical about AI’s ability to better human problem-solving, Jenner realised she could use the IP gained during her last eight years in business as a coach more effectively. 

In July 2023, she set up Actvo, a SaaS platform designed to help leaders activate latent talent in their teams, as a “tech extension” of START Now. 

To help make the AI dream a reality, Jenner needed someone with technical know-how to balance her experience. A few months later, Simon Plant came on as a fractional CTO. 

With his 20 years’ venture tech experience and time as a software engineer Plant, who has also previously worked in the AI space, was a perfect balance for Jenner and all her experience with humans.

“That’s when the real work began,” she says. 

Actvo went live in February 2024, having signed its first customer prior to launch. 

Jenner has others in the pipeline who she hopes to sign in the coming months. She is positive about AI’s potential in the professional coaching and mentoring space:

“I genuinely believe AI is going to be better at this than humans, so that’s the mission. We’re creating a scalable new way of supporting people in their careers.” 

What we’re doing doesn’t exist at the moment, says Jenner. 

‍

Jenner – with a few additions from Plant – shares how Actvo uses AI in a sit down with Caffeine. 

‍

How are you using AI at Actvo?

MJ: We’ve built our business with AI at the centre because we saw the opportunity to do something better than humans can, but having an AI business is like saying you’re an internet business. You have to get specific. 

There are four ways we use AI. 

The first and most obvious is creating digital personal guides – this is a big category opening up. It’s similar to the concept of assistants or co-pilots, but our AIs are trained to be experts in your personal career needs. They can’t tell you the weather or what to buy for dinner.

The second is personal recommendations. With AI, we create a superpower statement for individuals, crafting personalised statements and value propositions. What does that mean? It’s like, ‘What can you offer that is uniquely you?’

For example, this has helped one of our clients who’s a migrant. As English isn’t her first language, she struggled to articulate who she is and what she’s great at. By entering her data, the machine wrote it for her. 

Then, there’s predictive analysis. If we put it into a business, we can create lots of sentiment data or cluster employees together, providing predictive insights about employees to leaders, which companies value. 

The last is machine learning. We use supervised learning at the moment – I’m the main trainer – but we’re soon moving to self-enhanced learning, where the machine will basically teach itself. We’re not quite there, but it’s on the roadmap for the next 12 months. 

What has been your biggest challenge with AI?

MJ: The pace of it. It’s a continuous investment. There are no paradigms fully proven yet. Even the tech giants around the world can’t agree on terminology and where things will land. It is hard building into a sandstorm, you never know what’s coming. 

The second challenge is tuning the application to be as good as we want it to be. I believe it will be better than a human, but it’s like training a child. You’ve got to be thoughtful, constantly watch it, and keep up the pace. Right now, I supervise the machine all the time, but we hope to automate supervision, which will come with more funding. 

So that’s been a lesson; from the outside, it looks really easy, but it takes grafting. 

SP: There are a lot of [large language models] being developed and released. They all have different strengths and weaknesses. It’s a challenge to navigate the landscape and choose which models to use for which use case. 

We believe there is value in using multiple models to leverage them individually for the variables that will produce the best results for the specific problem or challenge at hand. We are starting to get a good feel for this.

‍

How do you manage privacy concerns?

MJ: Privacy is number one on our ethics list, otherwise we’d just erode trust. We choose to comply with the highest global standards: the European General Data Protection Regulation. At the heart of this product is the idea that when you engage with this platform, your responses are ironclad. 

SP: None of our user data is used outside our proprietary models, which will never change. We also always aspire to make it clear to our users what data we hold and for what purposes this is being used.

Are there any AI trends that you are closely monitoring?

MJ: Everything. All the time. It’s hard to keep up. But, I follow two things closely. 

First is Personal Assistants, training an AI or large language model [LLM] to replicate yourself to get anything you want done. 

The other is natural language processing (NLP). People have different modalities for communicating. 

For example, if I were to say to someone, ‘Describe what happiness at work looks like,’ they would naturally either want to draw something, say something, or write something. We want AI to be able to do this, too, switching between codes and taking in photos, speech or drawings where appropriate. That will place us more into that virtual human space. 

‍

What do you see as the main benefit of AI in your category?

MJ: We’re interested in how the workforce is changing and see two main drivers. 

One is the growth of neurodiversity, which is a big percentage of the workforce. Neurodiverse people don’t necessarily respond to traditional processing styles. 

The other is the growth of true diversity from migrant workers in large organisations in specific sectors. For example, our first customer is in hospitality, and I’d say 50 percent of their staff are non-English native speakers. Many cultures find it difficult to speak up in groups or open spaces, and that puts them at a disadvantage. 

These changes mean the way we’ve helped people with their careers in the past will soon be obsolete. We believe AI can change this career guidance space and, if done well, help overcome any personal biases people may have in the workplace. 

‍