AI beyond the ‘as-is’: updating the human operating system to unlock transformational value

“Leaders are under pressure today to develop an AI strategy and start to unlock potentially big benefits as a result. Yet, for leaders to ensure that these new AI capabilities will drive the outcomes desired, they need to upgrade the human dimensions of the organisation beyond just the technology.”

Today, AI is being discussed in corporate boardrooms 3.4x more frequently than cloud and 2.5x more frequently than transformation. The reason for hogging the digital agenda: the potentially massive size of the prize. Goldman Sachs estimates that a 30% profit increase over the next decade could come from leveraging AI capabilities to increase productivity by up to 1.5% annually across big business. This enthusiasm for AI at the top table means the pressure is on for digital leaders to deliver.

At the same time, we have seen big tech pouring in multi-billion dollar investments to accelerate their ability to service this demand. As Alphabet’s CEO puts it, this is a ‘once-in-a-generation opportunity’. To capitalise on the insane popularity of generative AI tools like ChatGPT over 2023 (think 100 million users in the first month since launch), big tech is building GenAI features directly into their enterprise product portfolios. Their aim: ease the integration into the digital enterprise estates of businesses around the world, and put this cutting-edge tech into the hands of any/every employee. At the time of writing, we see Microsoft’s Copilot almost fully integrated across their 365 application suite, and a new service called Copilot Pro, which lets users build their own custom ‘copilots’ using the technology.

For digital leaders today, the heat is on to integrate these cutting-edge AI capabilities into their technology estates fast, and in turn, start demonstrating those big productivity gains expected by the top table. There already are many options to buy in these technologies on the market today. However, there are many more unanswered questions on what will actually deliver the outcomes targeted by boards, and what can feel like very little time to explore them properly.

Seasoned digital leaders know that, in the end, they’ll always be judged by the outcomes achieved with technology, over just putting the technologies in place, so they need to move forward wisely. However, with one in six UK businesses already embracing one or more AI technologies digital leaders need to close the gap fast, but how should they approach the AI question?

The ‘as-is’ challenge with AI

When it comes to any digital transformation, organisational experts will say technology is only ever 20% of the equation, and behaviours like shiny toy syndrome (also known as buying cool bits of technology without a solid business case) can be a big reason for failure. Very rarely does tangible value come from just plugging-in and switching-on new technology across the organisation. And unfortunately for stretched digital leaders today, it is unlikely that AI will be any different in this regard.

A well-established approach to this challenge is to conduct an ‘as-is’ mapping of the current processes across the organisation and identify use cases where these new tools could be integrated. Though this approach can work when switching out like-for-like from different vendors, the barrier to effective adoption can feel like it increases exponentially as the gap between ‘as is’ and ‘to be’ gets bigger.

However, for most new technology we deploy into the fabric of our organisation today, we will only see our targeted productivity and efficiency gains if we manage to catalyse effective adoption of the technology AND effective utilisation of the technology, by employees. We could describe this in simple terms as using the right tools while using the tools right. So, what does this mean for digital leaders looking to deploy cutting-edge AI across their organisations?

If we take GenAI tools as an example, we are effectively asking employees to reimagine their ways of working, often inverting the established approach to tasks today. For example, using a tool like Microsoft’s Copilot, rather than writing an email to a client and doing a quick once over in Outlook, an account manager might now generate a complete draft of the email using a prompt then proofread and/or tweak the email instead.

Even in this simple example, we are changing the core skills that the employee needs to lean on to be effective and flipping the cognitive model of the task. So, depending on our target use cases and user behaviours, this likely means a triple whammy of skills training, change management and performance tracking costs added to the AI tab, before even counting the potential upside. Yet there is a bigger question that is rarely explored by leaders embracing AI. Using our email example: is thoughtfully editing a generated email actually more efficient than writing it, and is that generated email still at least as effective in achieving its purpose as before?

There isn’t a way to faithfully answer this question because there are likely to be as many answers as there are employees, due to each individual’s strengths and weaknesses in the context of this new bionic (human-machine) task. And when we multiply that problem by hundreds of interventions across thousands of employees, it is very hard to predict the macro impact on both productivity and efficacy.

We know the transformational outcomes we are aiming for, but the actual outcome of putting AI into our ‘as-is’ tasks, processes and operations depends on our organisation’s human operating system. The human operating system is the people, processes, structures, behaviours, and culture that make up our organisations, that harnesses our technology estate to unlock desirable actions and amplify overall value. We are likely opening Pandora’s box by deploying AI capabilities without creating a corresponding update to our human operating system.

Updating the human operating system

If we want to realise the value of AI across the organisation, we need to make an update to this human operating system at the same time as deploying new capabilities in the technology estate. This means we need to understand what it will take for meaningful adoption of these tools across the workforce, and at the same time catalyse effective ways of working that will predictably deliver the performance outcomes that we are targeting from the new technology.

1. How might we ensure meaningful adoption of our new AI capabilities?

Committing to business goals can motivate and engage employees. This means we’ll want to invest in getting buy-in on the reasons why we want to adopt these new technologies, before getting into the how or what. Around the topic of AI, employees likely show apathy, curiosity and trepidation in equal measure. And even those who use consumer GenAI unofficially, as soon as leaders start talking about ‘AI is coming’ the red flags go up and employees inevitably worry about their jobs. However, many positive outcomes can be targeted beyond just cost reduction through productivity/automation, for example, UK retailer Marks & Spencer targeted an 80% reduction in warehouse accidents with AI. We want to build an exam question that will resonate, so don’t be afraid to explore goals beyond efficiency!

Building a mandate for change starts by taking the challenge to the people, clearly communicating the outcome you are seeking to achieve from deploying AI can help build a foundation of trust across the organisation needed to overcome the inertia to change. This case for change can help to get everyone on board with the need to embrace this new world and actively participate in making it a reality.

2. How might we develop effective ways of working with AI?

To ensure that the new capabilities we deploy are used by employees in a way that is effective in achieving the goal we are targeting, we need to nurture employee behaviours that work towards this desired outcome. One approach to achieving this is building a holistic picture of what drivers (aka what is working today, creating value) and blockers (aka what is getting in the way, holding us back) exist across our organisation’s human operating system today, in the context of the target outcome.

We can build this holistic picture of the drivers and blockers to our target outcome by running a series of exploratory design thinking workshops with employees. Start by sharing the goal or exam question, then in smaller groups, discuss and brainstorm the drivers and blockers that exist today in trying to achieve the goal. If you already have ‘as-is’ processes defined, you can also get the groups to map their drivers and blockers against these processes for more focused brainstorming.

This mapping exercise gives you a rich landscape of opportunities to come up with ideas for AI-based interventions to either accelerate the drivers or overcome the blockers. Even better if you can get employees to envision and articulate these interventions themselves with brainstorming and canvas-type design thinking exercises in follow-on workshops or sessions.

3. How might we scale and evolve AI best practices across the org? 

Our organisation’s human operating system is not static. It is a dynamic, living system with often unpredictable outcomes to any change made. This means we need to embrace an approach to deploying AI that lets us quickly test a variety of ideas, validate our assumptions, and then scale up what works in driving towards our goal. There are many approaches to do this depending on the focus, scale and constraints of each organisation, but a continuous improvement approach with minimum viable governance is a good starting point to be able to rapidly pilot, validate and scale up successfully.

As the organisation starts to make effective use of our new AI capabilities, new ways of working will emerge from the bottom-up. Having an established forum or mechanism for employees to be able to share learnings, iterate ideas and develop best practices will help in furthering adoption and increasing efficacy across the organisation.

In conclusion

Leaders are under pressure today to develop an AI strategy and start to unlock potentially big benefits as a result. Yet, for leaders to ensure that these new AI capabilities will drive the outcomes desired, as well as avoid unexpected/unintended/unwanted results, they need to upgrade the human dimensions of the organisation beyond just the technology. Upgrading the human operating system requires having alignment on outcomes, an understanding of what drives/blocks that outcome today, and a continuous improvement mechanism to evaluate and scale up effective change.