This article is part of the Executive’s Playbook to AI at Scale series.

Introduction

With the release of new AI tools such as Gemini 1.5 and Sora, organisation leaders today are rightfully excited about the potential of AI, even more so with Generative AI, to help achieve sustainable competitive advantage and move ahead with speed on renewed digital transformation journeys to accelerate growth. Yet many teams to this day are approaching the adoption of this technology without a unified value-driven implementation plan that harnesses collective effort from the entire organisation — beyond simply engaging in pilots that build models in silos for one-off use cases, all without a focus to scale AI for maximum value.

Modern business leaders, who are privy to the fact that AI needs to be scaled, will immediately identify this approach as a major roadblock to achieving economies of scale and ensuring the reusability of AI as a core business capability. A lack of awareness would lead to inefficiencies of capital investments as well as significant levels of wasted productivity as teams would often start such frantic development efforts from scratch and prepare new data every time a new model is developed.

This challenge isn’t exactly new, with the likes of cloud and big data analytics topping the list of transformation imperatives in recent periods — yet it is a “gold-rush” type of situation that holds enormous potential for agile and forward-looking companies to get right. Without a well-informed leader to champion a unified approach to adopting AI, organisations put themselves at a significant competitive disadvantage. And despite many turning their attention towards CTOs to lead this effort, it is CEOs who must recognise their role in setting a strategic plan for their company to build, deploy and manage AI with a sense of speed, efficiency and security.

Companies that do this well often have teams operating in sync with one another, aligned by a common understanding of the potential of the technology, enabled by a fluid way of working and guided by first principles set by the leadership to ensure AI development and management efforts are in line with long-term company objectives.

Moving from an era of experimentation and exploring POCs to now maximising the business value of AI requires large organisations to embrace the crucial cultural and organisational shifts and invest in building strong company foundations that when integrated with the new technology and tools pursue an end-to-end capability transformation rather than one that chases after siloed, bespoke solutions.

Today, global companies see AI as a must, not a nice to have — rightfully so, as the technology has matured significantly and customers demand to be wowed by increasingly personalised and seamless experiences driven by AI-powered interactions. The talent pool, technology platforms, data tools and services available have also evolved in numbers, giving any organisation the capacity to address the entire AI life cycle, enabled by an increasingly popular MLOps approach for stitching AI delivery capabilities together. However, companies still need aspirational yet hands-on leaders who can play a critical role in translating this blueprint into reality.

This article will serve as a guide for CEOs looking to orchestrate successful company transformation efforts to leapfrog and disrupt their respective industries by highlighting best leadership practices to scale AI and identify the right areas in the organisation to lead to successfully infuse AI into business processes, workflows and operations and ultimately, put in place a holistic plan to successfully implement AI at scale.

Align the leadership

Misalignment among leadership in technology comprehension and strategic planning will almost always lead to disaster — scattered execution of a company’s transformation — and with technology as complex as AI, plans of this magnitude can make or break the core business as AI often involve and have lasting consequences on many different areas of the business.

CEOs should spend enough time establishing a common understanding among the C-suite about the very nature of the technology and the potential value it can contribute to business metrics. The leadership team should have a contextual understanding of AI, or even better be well-versed with the associated analytical capabilities and key components that make up the technology. Put in place a series of knowledge programs to learn about the importance and types of foundational models, training architectures, general capabilities of LLMs, hardware requirements, as well as the surrounding ecosystem and parts of the value chain that enable AI systems to work.

Once the leadership is on the same page when it comes to AI knowledge, they can branch out to explore the latest developments in AI, relevant innovation trends and best practices from leading companies inside and outside their industry, who are further along the AI journey to learn where and how to play — before developing a shared vision on how AI will disrupt their own organisation and transform day-to-day operations. CEOs must crucially step in as their leaders brainstorm and form directional thoughts around the intrinsic value of AI and facilitate a long-term focused discussion that captures and drives the company’s future value rather than looking at how AI will simply boost performance and valuation in the next earnings call.

One way to do this is by identifying key domains to pursue that will add long-term value or transform business domains end-to-end. CEOs should avoid starting too small as implementing a set of uncoordinated use cases will produce little returns. Instead, find the groups of high-value use cases by identifying wider, well-crafted domains (customer journey, opportunity management, product development), where both current and future use cases across business functions can produce the most tangible outputs by leveraging similar data assets.

CEOs should also be the first ones to raise their hand and question extensively the existing readiness and maturity of their organisation’s cross-functional delivery capabilities needed to fulfil ambitious AI transformation plans. Capturing value from AI requires the leadership to be crystal clear on seven key drivers that will determine the success or failure of any efforts around building a production-line environment for scaling AI — vision, data, talent, value, operating model, governance and system.

CEOs can ask themselves and the leadership team the following initial set of questions to get started on creating a common understanding of where they are in this journey. Consequently, identify the gaps that need to be addressed before running at speed towards fully integrating AI into the business.

AI readiness questions for CEOs and leadership

With improved clarity around organisation readiness, further effort will be needed from the CEO to build executive cohesion on what are the right set of AI ambitions for the company and take the time to solicit explicit agreement on individual leadership commitment towards driving business outcomes that are measurable with clear KPI improvements. CEOs should leverage these metrics in long-term planning and form a type of pact with the entire C-suite that spurs follow-through actions and ensures that execution is on-point.

Empower and organise for scale

At the core of every digital transformation is the transformation of people. The evolution of how they work together is what fundamentally sets the flywheel of change in motion. These efforts very rarely work without supporting and enabling the right set of people to drive a common agenda from top to bottom. If those qualified personnel do not yet exist in the organisation, there is no better person than the CEO to kickstart the process of putting a strong digital talent bench together.

A good starting point is the team that works most closely with the highest-ranking executive in the company; identifying gaps in leadership positions that are best filled with AI-experienced personnel that can shape the organisation in multiple ways — establishing technical standards, setting the tone for ways of working and driving organisational change initiatives. Commit time to conduct a broad search of potential additional senior hires with strong digital backgrounds such as in data analytics or machine learning.

Another option is to set up a cross-functional AI council that will not just oversee and steer the implementation but identify which part of the strategy will be affected by new applications, develop a use case pipeline, and decide how to allocate finite resources to achieve success with AI. While many CEOs will rush to assemble a team represented by each function, they may end up with too many hands at the wheel. Instead, set up a small working group at the helm which is supported by an extended team that can move fast and get tasks done with the help of the larger organisation. However, certain key functions are a must-have in the council composition, which includes IT, strategy, customer, creatives, HR, risk and compliance.

CEOs should also manage their key stakeholders during times of change, particularly the board, by establishing an advisory committee consisting of internal council members, external experts and industry leaders. This hybrid committee will advise and provide insights on the latest AI developments and support strategic thinking about the long-term implications of AI, especially when it comes to ethical responsibilities and usage of AI.

To drive this transformation, changes made at the top must also be reflected structurally at the lower levels; starting with flattening hierarchies and encouraging the germination of cross-functional, agile teams to bring business, technology and operations closer together into a working model that speeds up experimentation and learning cycles. These sprint units should have the multidisciplinary skills required to develop AI applications that directly impact a customer experience or focus on creating reusable services for functions in the company.

CEOs have a few options here; establish a separate operating unit, designed and structured with sprint units, which ensures complete independence in building digital solutions; move to a product and platform model, which essentially reapplies the concept in the core business or extend the sprint unit working model to the entire business — including sales, marketing, R&D and other functions that would find value from working in small, cross-functional teams.

CEOs, at times, regarded as the Chief Culture Officer, also act as key role models and champions of organisational culture and mindset change that often dictate whether AI transformations succeed or fail. At the pace of how AI is disrupting operations, there is no question that the concept of “fail fast, learn fast” becomes a cornerstone to success in scaling AI. Teams need to shift from slow, rigid and risk-averse to adaptable, operating with a learning mindset and being open to pulling the plug on something that isn’t working and moving on to the next idea “quickly”. CEOs can set the tone for this culture, putting in place working processes and management frameworks that encouragement of innovation, risk-taking, and experimentation to create minimum viable products in weeks rather than months — letting employees know that it’s OK to try and fail.

• Piyush Gupta from DBS, one of the World’s Best Digital Banks, describes his earliest AI projects not as failures but as “signalling tools” for the organisation that prompted a culture of experimentation and overall learning about the different types of AI projects that would be most valuable to the bank. With the foundation and capabilities to run hundreds of experiments involving data, AI and machine learning in a single year, DBS is now light-years ahead of its competitors in terms of AI and digitalisation.

Set the right aspiration

Getting all internal stakeholders to support this transformation is beyond crucial. CEOs can rally together the level of camaraderie needed for this change from the board, investors and especially employees. Start by breaking down legacy organisational constraints to reduce the inertia and gather attention around a strategic vision. Top CEOs set up such aspirations by thinking about obligations to shareholders, the relative strength and purpose of the company, business value drivers, innovation opportunities, future trends as well as personal goals. This strategic vision would likely be more effective when reframed and CEOs create a reference point for what success will mean.

Going back to the DBS example, key stakeholders rallied around the mnemonic “GANDALF”, from Lord of the Rings), which represented the leaders of the technology market as each letter represents a different company — G for Google, A for Amazon, N for Netflix, F for Facebook and so on. Initially, the letter D in the middle, seem to not make much sense, but in fact, it was reframed into a call to action for stakeholders to realise the ambitious vision for DBS to join the big leagues of iconic tech leaders. Such a reframing helps internal stakeholders acknowledge that the company competes on a bigger stage than their industry and they are each working towards something truly aspirational.

CEOs can accelerate the change management process by investing quality time in the early stages to communicate clear expectations of different teams — from the leadership to middle managers and other key change functions, which is to generate sustainable value from developing, deploying and maintaining AI at scale. While the vision sets a directional north star, CEOs should make it clear that everyone sees AI applications and services as equal to other business-critical systems and must be fully utilised to drive business value across the organisation.

It also helps that the leadership address how all workers will fit into a new AI-oriented culture and dive even deeper to identify and clarify the value drivers of AI for different teams and function leaders. CEOs should go beyond the typical strategic address at town halls to help people at every level of the organisation understand the nature of such transformations and why they should care about AI in the first place.

To further galvanise and reinforce this transformative process, it is in the best interest of organisations to regularly track their change progress with pervasive performance measurement metrics that span across all levels of the organisation, even the leadership team. Comprehensive success measures that can be quantified with specific key performance indicators provide useful evidence for CEOs to understand whether transformation efforts towards becoming an AI-enabled company are working and provide a collective signal so teams can course correct as needed. While operating models help ensure shared goals and joint accountability over the development and deployment of AI, establishing incentive-baked performance yardsticks will not only bring clarity to individual roles but improve employee engagement and foster healthy environments for professional and personal growth.

Example measures to assess AI transformation efforts

As AI proves its worth beyond mere hype for experimentation and is starting to show clear signs of business value, involvement from the leadership is crucial to spearheading the organisation into a bright, AI-driven future. With the demand for every company to be an AI company, the pressure on CEOs is mounting. Leadership alignment, growth of sprint units, a learning mindset and a unifying purpose can lay strong foundations to put into motion the shift towards a systematic and cohesive approach to the development and management of AI.

Such fertile grounds to cultivate a make-or-break type of transformation are of utmost importance to create immediate value. What is even better is that it sets up the right foundation for organisations to benefit from the innovation flywheel of more collaborative working styles between interdisciplinary teams, flatter hierarchies, greater ownership of outcomes and integrated domain-focused capabilities in the long run — the journey to successfully scaling AI begins with the CEO at the forefront leading the charge.