Is AI the best career coach? Experts say it’s complicated

By Rosalyn Page

A well-configured artificial intelligence (AI) careers adviser is like having a brilliant, knowledgeable colleague who is available 24/7. But creating an effective AI coach takes time, says Taleist Agency founder Steven Lewis, and some aspects of coaching will always be better delivered by humans.

The key to getting meaningful insights from an AI coach is taking the time to train it with relevant information.

“AI can reflect what you know and where you are thin, as long as you have fed it the right information about your context and goals,” Lewis says.

Therese Rahme, principal consultant at Sydney Career Coaching, says one of the most direct ways AI can act as a careers adviser is by performing benchmarking and skills analysis. However, like any AI tool, its effectiveness depends on robust prompt engineering and the considerable thought that users put into their instructions.

“If you want to put your entire skill set in there, for example, what would that look like?” Rahme asks. “Are you more human-focused, skills-focused or innovation-focused?”

Knowing when to question AI is just as important as knowing what questions to ask, Rahme says. She cautions people to slow down and consider their options, rather than automatically assuming the AI tool is providing the right advice. The speed of AI can create false confidence, she advises.

“It is generated so quickly that people assume it is right, especially when they are already feeling overwhelmed or uncertain.” 

The decisions an AI coach makes do not need to be completely autonomous, because users can retain control over the decision-making process with the right inputs.

“You can feed it source material such as reports on decision-making biases and then ask it to incorporate this into future sessions to help challenge your thinking process,” Lewis says.

Where human career mentors outperform AI

While an AI assistant can help optimise a résumé, the risk is that the output could end up sounding generic or formulaic. 

A further limitation is that when AI is used to research roles, it can only access information that is publicly available, not the specific data held by job boards.

Rahme says that human career coaching draws from both spoken and unspoken information, but AI cannot pick up on the emotional cues or other interpersonal signals in human-to-human interactions.

“It is not factoring in ambiguity, identity or the emotional nuances that come out in a coaching conversation,” she says. “It cannot recognise the organisational politics or even the confidence issues that might be there. We have to look into burnout, values conflict and so on.” 

In addition, AI is unable to offer personalised, human-oriented insight and support, Rahme continues. It misses the individual context.

“So much of human communication goes beyond words alone: tone, gestures, pauses, emotional cues and context all matter in a coaching conversation.” 

AI also has a confirmation bias baked in. “It has a terrible habit of telling you, ‘this is great, here are 10 ways to get it done’. But it is not always that simple,” she says.

While people may rely on AI to complete job applications, they risk missing important details and context that only the applicant knows.

“Is that real certainty we can sit with and grow from? Or is it just ‘it can populate this for me, so less for me to think about’?” she says.

How to build an AI coach responsibly

Lewis recommends developing an agent to act as an adviser by providing professional knowledge, client context, communication style, personal domain expertise and skills such as financial analysis frameworks. This can extend to tools that support technical upskilling and performance benchmarking for accounting and finance professionals.

However, expert guidance is critical to establish the security, privacy and governance requirements of an organisation-mandated AI tool.

“Just as firms would not build their own calculators, most accounting partners should not be the ones building their own AI tools,” Lewis says. “Bring in someone who knows how to build it properly.”

Lewis sees the AI adviser relationship as mirroring good mentoring: the professional brings context, and the AI brings challenge and structured thinking. Neither defers entirely to the other.

 

_{Production credits
Banner image saifulasmee chede via Getty Images}