Business storytelling expert and bestselling author of Business Storytelling from Hype to Hack, Jyoti Guptara, outlines how pioneering AI in organisations can boost productivity within businesses.
Millions of professionals have started using AI in organisations to boost personal productivity.
However, can AI help change culture, implement strategy, and make us more human at work? The next development for organisations will deploy AI for organisational effectiveness.
The world’s leading executive coach patiently listens to me. Probes. Suggests. For hours, if I want. Day or night. Marshall Goldsmith is a trusted advisor for top executives, and now he’s helping me take my business – and life – to the next level.
Why do I have this kind of access to one of the most in-demand people on the planet? Free of charge? Because the Marshall Goldsmith I’m talking to is not the real-life individual but his AI alter-ego.
For the last four months, Mr Goldsmith has been training a bot, which he plans to replace him. Feeding the text generator – technically, the large language model (LLM) – his 48 books was an easy first step.
Getting the AI to take on Marshall’s conversational personality requires more work. But how much time before the above scenario becomes a reality? According to current indications, not long.
Artificial encouragement
“After three months,” Marshall told me at a session on AI I recently convened for the Global Peter Drucker Forum, “Generation 3 of MarshallBOT can answer questions given to me, I would say in a more articulate way than I can, about 50% of the time.
“This is a primitive version … MarshallBOT 22 I have no doubt will be able to answer 95% of the questions I get better than I can – as judged by me.”
Goldsmith plans to give away all his knowledge and experience of pioneering AI in organisations for the public benefit. If a generation of thought leaders are inspired to give away everything they know, what will this mean for coaching? For learning and development as a whole?
Pioneering AI in organisations takes businesses beyond efficiency
The initial allure of pioneering AI in organisations was its promise of efficiency. Automating routine tasks and reducing human error were the primary selling points. Increasingly, AI is about effectiveness rather than just efficiency.
I work with companies shifting their perception of LLMs from anonymous servants to partners – or coaches. This is also how cutting-edge models are being designed.
For example, the AI startup Astrid Education began as a language training coach. When Astrid needed to hire more programmers and was inundated with CVs, they used their technology for a second purpose, screening applicants. Issuing automatic speaking tests to candidates cut down initial interviews by 80%.
Now, a third application of their foundational model is the optimisation of customer service or sales calls, with Astrid’s bot listening in and providing feedback.
Who owns the data?
All this raises the question of IP, data ownership, and privacy. As a storytelling expert, I’m hesitant to share my formulas publicly, knowing they’ll be harvested. When employees asked ChatGPT to improve sensitive code, Samsung accidentally leaked intellectual property, leading some innovative companies to debate whether employees should be allowed to use public AI models.
Andreas Kempf, a ZEISS global leadership team member, explained their cautious approach to AI adoption to me this way: “Pioneering AI in organisations is a great opportunity – to lose your proprietary knowledge.”
One solution is to train local, in-house LLMs on an internal corpus of information. This can then save countless hours in knowledge transfer. Think back to your first few weeks in a new job.
Like me, you may have had questions you never asked because you didn’t want to interrupt busy colleagues – even though you needed those answers to move forward. Chatbots in companies with high turnover are especially useful for saving onboarding costs and helping newer workers shorten the experience curve.
Chatbots can be equally valuable for established employees by making knowledge available. Companies like the Swiss bank UBS are already pioneering AI in organisations to help managers navigate their career paths. Unblocking the ‘traffic jam’ for anyone who wishes to tap the knowledge of a senior colleague is an example of the kind of bottleneck AI could help free once and for all.
From personal assistant to workplace personality
In August 2022, the publicly traded Hong Kong firm NetDragon Websoft replaced its CEO with an AI-powered ‘virtual humanoid robot’, Tang Yu. Reportedly, many employees are more willing to express their demands or convey their suggestions to the digital CEO.
Which company will be the first to immortalise its iconic founder’s ‘conversational personality’ to ensure long-term commitment to its founding purpose? Will it become normal to interact with virtual manifestations of brand personalities? Whether this makes decisions more or less transparent will depend on how they are implemented. Just as there are different brand personalities and leadership styles, other AI models will have their own styles, too, by default or design. Some brands may choose to have a more authoritative AI model.
In any case, LLMs hold huge potential to save us time. In the form of enterprise search, some departments now save two to three hours on answering daily questions per person, freeing up highly qualified employees. At the same time, the inquirer gets an immediate answer. But what if the true potential of an LLM lies not merely in retrieving and remixing the data it was trained on?
Introducing the corporate moderator
AI tools like Humu analyse employee surveys, then send ‘nudges’ – texts with behavioural recommendations – to managers and employees. But one can’t dialogue with a poll or a nudge. An LLM could facilitate two-way conversations between various roles, departments, or stakeholders.
What will this kind of real-time analytics mean for the discipline of management? When it comes to human-centric management, our recent record is startlingly poor.
Gallup CEO Jon Clifton, author of the book Blind Spot, told me how leaders missed the global rise of unhappiness in workplaces and beyond. Would this have happened if pioneering AI in organisations was being used to its full potential? An AI model could have brought this to our attention – but only if programmed to do so (which would indicate the topic was not a ‘blind spot’).
LLMs have biases and blind spots, too, a reflection of the data on which they were trained. Worse, they often gloss over these with ‘confabulations’ (plausible but fabricated responses). Perhaps we should be relieved. Decisions will require a human-AI partnership, meaning humans are not obsolete.
Pioneering AI in organisations: It’s over to humanity
I hate the idea that soon AI will be able to do nearly all skilled tasks, and I don’t dismiss dystopian concerns. But everyone concerned about pioneering AI should be at the forefront of its implementation.
Then you, like me, can be a part of the conversation – taken seriously. Crying foul from the sidelines will only mean becoming sidelined as Luddites. The more we understand and use AI, the more ownership we’ll have and the more we can exert control over its direction.
One of the impressive – and alarming – characteristics of machine learning is how, after endless experimentation, they can immediately share their progress. When one machine masters a skill, such as catching a ball, it can all suddenly catch it. We need to work together in the same way. The future can indeed be an age of abundance, but it will require co-operation and a good common narrative. Isn’t that something we all need some coaching on?
