Organisational AI & The Future Of AI Operations

Hello and welcome to the latest edition of AI Strategy Brief. In this edition we’ll review some recent industry developments in AI and consider the future of AI Operations as a potential business function.

?

AI Industry Updates:

• Software – Microsoft & OpenAI On Next Steps
• Computing – Intel Develops Fastest AI System Yet
• Chip Manufacture – AI Demand Strong As Nvidia Revenues Soar

?

Microsoft & OpenAI On Next Steps

At Microsoft's annual developer conference, Microsoft Build, OpenAI CEO Sam Altman made a surprise appearance to discuss recent AI advancements and future expectations. Joining Microsoft CTO Kevin Scott on stage, Altman expressed amazement at how rapidly AI technology has been adopted, highlighting the significance of APIs in this growth. He emphasized that future AI models will become smarter and more useful, noting improvements seen from GPT-3 to GPT-4, and mentioned upcoming enhancements in speed, cost, and multimodality, like the new GPT-4o with Voice Mode. Altman advised developers to seize the current momentum in AI, comparing it to transformative periods like the advent of mobile phones and the internet. Microsoft also announced the general availability of GPT-4o in Azure AI, further leveraging OpenAI's technology to enhance its services.

This comes at a time when Microsoft is gearing up for an AI-driven future with significant investments and technological advancements. Recently, CTO Kevin Scott emphasized the transformative potential of AI to generate societal benefits globally. To support this vision, Microsoft announced a $2.9 billion investment to enhance AI and cloud infrastructure in Japan, aiming to train three million people in AI skills over the next three years and establish a new Microsoft Research Asia lab in Tokyo. This follows on from similar announcements on developing AI skills in India. Additionally, Microsoft has launched AI-powered devices, such as the Surface Pro 10 and Surface Laptop 6, featuring Intel Core Ultra processors optimized for AI workloads. These devices integrate AI capabilities, including Windows Studio Effects for video conferencing and AI-enhanced dictation in Microsoft 365 apps, designed to boost productivity and creativity for knowledge workers.

?

Intel Develops Fastest AI System Yet

Intel, in collaboration with Argonne National Laboratory and Hewlett Packard Enterprise (HPE), has announced that its Aurora supercomputer has surpassed the exascale computing threshold with speeds of 1.012 exaflops, making it the fastest AI-focused system. Designed as an AI-centric supercomputer, Aurora is already facilitating groundbreaking research such as mapping the human brain's 80 billion neurons and accelerating drug design through machine learning. Equipped with 21,248 Intel Xeon CPU Max Series processors and 63,744 Intel Data Centre GPU Max Series units, Aurora represents the largest GPU cluster worldwide, leveraging Intel’s innovative Xe GPU architecture and software tools to enhance scalability and developer flexibility.

?

The implications for chip and hardware development around AI are clear. Aurora’s success underscores the growing importance of advanced GPUs and optimized architectures in handling complex AI and high-performance computing (HPC) tasks. This drives the demand for more efficient, powerful, and scalable hardware solutions. The expansion of Intel’s Tiber Developer Cloud and the deployment of new supercomputers integrating Intel technologies highlight the need for robust hardware to support large-scale AI model evaluation, innovation, and optimization, further pushing the boundaries of what AI and HPC can achieve.

?

AI Demand Strong As Nvidia Revenues Soar

Nvidia’s revenue soared 262% in the past quarter, driven by record demand for AI chips, surpassing high expectations. CEO Jensen Huang announced the company's new Blackwell chips will contribute significantly to revenue this year, continuing their trend of releasing powerful new chips annually. Nvidia's AI data centre revenue jumped 427% year-on-year, with shipments of Blackwell chips starting soon. This surge in demand from tech giants like Google, Microsoft, Meta, and Amazon has pushed Nvidia's market value to $2.3 trillion, making it the third-most valuable US-listed company. The company also announced a 10-for-1 stock split and a 150% increase in its quarterly dividend.

?

The implications for AI chip and hardware development are substantial. Nvidia's rapid advancements and substantial market gains highlight the growing importance of high-performance, AI-optimized hardware. This success pressures competitors like AMD and Intel to innovate and keep pace. The ongoing high demand for AI computing power will likely drive further investments in specialized AI chips, enhancing performance and efficiency in AI applications. This competitive environment fosters accelerated development and deployment of cutting-edge AI technologies across various industries.

AI Transformation Blueprint

Today’s newsletter is sponsored by Anordea’s AI Transformation Blueprint. At Anordea, we advise CEOs and senior leaders across Financial Services, helping them tackle their toughest problems. With the rapid pace of development and change in the AI space, many leaders are faced with a paradox of how to build on and govern the technology that is as complex as it is powerful.

?

Our AI Transformation Blueprint solves this paradox quickly and effectively by giving leaders immediate insight into the state of AI across their organisation and industry, highlighting the opportunities and challenges which need to be addressed to maximise their AI opportunity in a safe and well-governed manner. Based on our leading expertise and ongoing research in the AI Strategy and Governance field, the AI Transformation Blueprint offers a fast and high impact solution to one of the biggest challenges facing FS leaders today.

?

Working with a dedicated member of our senior leadership team, the benefits of this process include ROIs in the strategy, governance, operational, financial and personal arenas. For those interested in exploring what our AI Transformation Blueprint involves, and how it can help your organisation gain a leading advantage over competitors, reach out to one of our senior partners here or contact us at [email protected].

Organisational AI & The Future Of AI Operations

One of the things I've been tracking over recent months, as part of both my advisory and research work, is the manner in which artificial intelligence is being used actively in organisations. While advances continue in frontier models such as GPT-4o, the application of these technologies in frontline organisational operations has advanced at a much slower pace. This relates to the big challenge facing AI use across industries today - how can AI tools be effectively used while mitigating the challenges of managing them?

?

The Ultimate Centaur Challenge

When we look at an organisation and how artificial intelligence tools can be deployed across it, in some ways we start to see the ultimate Centaur challenge emerge. While Centaur usage is often unique to specific use cases and contexts, and we are still grappling with the myriad of opportunities to use these technologies in new ways, a continuing challenge relates to how they are coordinated at an organisational level. This is the major challenge facing organisations today and probably for the coming decade in terms of technology use and development. With the explosion of AI tools and use cases across organisations, and the related advantages that come from using these tools, there is a clear need for businesses to engage and deploy AI. However, what is lacking is a clear framework and system for managing the deployment, development, and use of AI tools across enterprises. In many ways this represents the ultimate Centaur challenge given the complexity of the coordination task. However, as yet, there is no integrated AI technology to help leaders coordinate enterprise level efforts.

?

AI’s Not Suited For Organisational Support

An interesting aspect of the AI tools we have seen emerging, that has probably not occurred to most, is the nature in which they are orientated towards the individual level rather than the group level. The AI tools we now find proliferating across industries and consumer offerings, are all designed to aid an individual. However, these technologies offer little in terms of support at the organisational level. This differs from most technology advancements seen in previous decades, which have been focused on creating and delivering value at the organisational level. In some ways the AI technologies we have seen emerging have developed in a way that breaks the traditional digital innovation mould. Perhaps the most disruptive aspect of AI is how it is oriented to support an individual very well, but organisations less so. As of yet, AI tools do not help solve organisational level problems.

?

The AI Value Proposition Vector

To illustrate this point, we need only to look back at the other technologies that have developed and proliferated across industries over the previous two decades, particularly in terms of information technologies that support knowledge work. If we take an example of something like database software for instance, this gives a clear example of how technology or software emerged in a way that offered great value at the organisational level, and also increased coordination capabilities. A more recent example would be the proliferation of cloud-based project management software, running from the original Microsoft project, through the Kanban tools we see available in the cloud today, and even those integrated with wider software packages such as Office 365. Again, these examples all highlight tools that help solve organisational level problems while creating great value in terms of the coordination of group level efforts. This is in direct contrast to the AI tools we see emerging, which are orientated almost exclusively towards the individual usage level and perhaps operate in a way that is shrouded from the gaze of the organisational level, or worse, even confounding to it.

?

The Potential For AI-Based Advantage

Drawing from the old strategic management tool kit, it's useful to consider the VRIO framework here, which is one of the most basic and well-known frameworks in strategic management. The VRIO framework proposes that any resource an organisation has can be assessed in terms of its value, rarity, inimitability, and organisational support. The purpose in assessing resources along these lines is to identify those which can be a source of competitive advantage for the enterprise. If we run AI tools such as todays GPT models through this framework, it gives us an interesting result. Are these AI tools valuable? Yes. Are these AI tools rare? Here we would have to say no, given the mass access that has been opened up to them. Perhaps what is rare is their application and integration into business operations. Are AI tools difficult to imitate? This depends in terms of how they are used. The platform level technology that AI presents on the face of it can be deployed in so many ways, that it is hard to tell what is inimitable and what is not. And finally do these AI tools have organisational level support? This is clearly where the advantages that can come from AI tools are most inhibited. There is little organisational support for AI tool usage across industries at this point in time, for a range of practical, technical, governance, and ethics reasons. However, the AI tools themselves are perplexing in how they make it difficult for organisational support to exist. This is the nature of the governance problem we see in organisations today that are interested in using AI tools. If AI is to deliver value for organisations and realize the promise of its inherent value propositions, we need to solve each of the areas of the VRIO question, beginning with how we can create organisational support for AI.

?

Second Order AI & The Future Of AI Operations

This brings us to the topic of AI operations, and what I would call second order artificial intelligence usage. At the moment the AI technologies we see being deployed are first level usage in that they support Centaur usage in frontline activities. What we need is the development of AI tools at the second order level which are designed to coordinate AI tool usage across organisations. Given the complexity and the diversity of AI deployments across a large organisation, the multitude of use cases possible, and the speed at which AI develops and evolves, it is likely that this coordination task would be incredibly difficult for even a team of well experienced and resourced veteran staff. I propose that this is the ultimate use case for Centaur deployment in an organisation and that over time we will see two levels of AI usage emerge within an organisation, at the frontline level and at the coordination level. Without this second layer of AI coordination, it is difficult to see how meaningful organisational support for AI can exist at the enterprise level.

Could AI Operations Be A New Business Function?

So how might this second order coordinating AI work, function, and what would it look like? To answer this let’s look at the realm of data analytics which is already making great strides in supporting decision making across the organisation. Data analytics has been bringing together and harnessing data from across the organisation, often deployed in centralized dashboards, for some time. Is it so fanciful to imagine that these dashboards will either evolve into or be replaced by some sort of integrated AI tool? I would suggest this is a natural evolution we will see over the next 5 to 10 years. However, I think there is a need to go further than this. Just as the value of having a PMO should be clear to any large organisation engaged in significant project work, I think the value proposition of having AI Operations explicitly structured in a corporate schema should now be clear. This will probably likely fall under the aegis of a Chief AI Officer. However, I think there could be a need for more than something on a PMO level, and it is possible we could see AI Operations emerge as an entirely new business function in and of itself. This stems from the nature of AI usage across the organisation and how it will become embedded in everything and every part of what the organisation does. Compare this to human resource management or finance as an example. People are embedded across the organisation, as is the flow of funds. And yet there is great value in having a centralized management function for each of these core areas of an organisation. I think in the future we will see AI usage become something similar and require perhaps the same treatment that is given to the human and financial resources of an organisation. This is the business case for AI Operations as a standalone business function.

Leadership Takeaways:

• Emerging Business Function: AI operations could evolve into a new business function, similar to human resources or finance, given its potential to be embedded in every part of the organisation.
• Need for Coordination: Effective AI deployment across an organisation requires a clear framework and system for managing AI tools, highlighting the complexity of coordination tasks.
• Individual vs. Organisational Support: Current AI tools are primarily designed for individual use rather than addressing organisational-level problems. This gap needs to be bridged to leverage AI's full potential.
• The Ultimate Centaur Challenge: Coordinating the use of AI tools at an organisational level remains a major challenge due to the complexity and diversity of AI applications.
• Unlocking AI Value: While AI tools are valuable and increasingly common, their organisational support is lacking, which inhibits potential competitive advantages.
• Second-Order AI: Developing AI tools at a second-order level, aimed at coordinating AI usage across the organisation, is essential. This approach would address the current limitations of AI tools in solving organisational-level problems.
• Potential for Centralized Management: AI operations might require centralized management similar to project management offices (PMOs) to ensure effective coordination and utilization of AI across various organisational functions.
• Chief AI Officer Role: The establishment of a Chief AI Officer could help drive the integration and coordination of AI tools, ensuring they deliver maximum value at the enterprise level.

That’s it for this edition of AI Strategy Brief. Subscribe now, and if you like what you read today, please like and share it with your network to help me reach a wider audience. Have a great day, and I'll see you next time!