Not your father’s avatar: The real future of artificial intelligence

Generative artificial intelligence (genAI) is the latest milestone in the “AAA” journey, which began with the automation of the mundane, lead to augmentation — mostly machine-driven but lately also expanding into human augmentation — and has built up to artificial intelligence. Artificial? What if business users started developing alter-ego avatars that performed real business functions? If this is one aspect of the future of AI, what will be the implications for technology leaders? 

In February 2024, Bjorn Jesch, CIO at German asset management and investment firm DWS, posted a video of his own AI-built avatar providing a glimpse of how he saw the market developing in the near- to mid-term. While avatars are by no means, a new phenomenon — many abound in the esoteric worlds of gaming, sci-fi and film, as well as in learning and education — Jesch’s post is, perhaps, among the first of a senior business executive’s use of a genAI alter-ego for personal and business benefit in the course of daily work. Personal because it made life a tad less stressful for a rushed executive — impossible as it is for someone like him to be in several places at once. And the business benefit, of course, resulted from a timely deliverable.  

But then, Jesch is no ordinary CIO — he is the ‘chief investment officer’ at DWS, not the same kind of CIO that acronym usually refers to! Since that first experiment, Jesch’s avatar now regularly podcasts without him recording anything; instead, an AI script extracts relevant information from various sources including meeting notes, local files, public sources and data stored in DWS…and then feeds it to the avatar which has now ‘learned the style.’ 

In the works is a desktop version that could attend virtual meetings. Think of this as a moonshot of certain aspects of genAI that might ultimately lead to digital-human convergence. The implications of this convergence are many and there are already hundreds of instances in business, software, devops, agentic processes and many more sectors with similar stories and use cases. From a CIO’s perspective, there are three key points to note. 

Beyond the CIO 

First, no formal request was made to the CIO — the chief information officer — to execute a project, nor were business requirements defined, or even an elaborate budget needed. The business user was able to find the tools necessary to realise an objective without any change, integration or impact on any legacy systems. 

Second, users were able to seek and procure the tools that could deliver an objective independently. These tools were mostly usable with minimal concern about underlying technology and its longevity, TCO or quality — possibly because the barriers to both entry and exit were deemed insignificant. 

Third, and most importantly, the AI avatar became an extension of the business user — a human worker who became digitally augmented and increased the productivity of that individual. Effectively, a self-learning, digital twin was born. Even though still in its infancy, with age it will gain the maturity to independently execute many tasks that will be delivered in an increasingly autonomous fashion, while all the time preserving the integrity of, for instance, its owner’s mannerisms, speech and nuanced uniqueness. 

About two years into the genAI journey and about 20 years in for AAA, 2024 is the year in which the convergence of the digital and human worker is truly beginning. If this is indeed the launch pad, how will multitudes of proliferating business avatars of real individuals — not fabricated talking heads — mature with time? 

AI digital workers…at scale

Aside from the business impacts of cost, efficiency and productivity that are becoming a familiar refrain in any analysis of genAI’s future, its lateral influences are less-common subjects of study. The expectation that some manner of convergence of digital and human staff in most corporate settings will grow exponentially is accepted, even though not all of this relates to digital ‘versions’ of real executives and employees. Enterprise CIOs, CEOs and, indeed, CHROs/chief people officers will be less prepared to deal with the many millions of executive avatars that could be performing on office screens, bringing valuable content, opinions and decisions to bear, driven by the small language models of their human owners.  

How do we measure, appraise and reward these converged individuals? How will the human capital management (HCM) systems deal with ‘them’? And when the inevitable resignation hits the boardroom table, will their avatars follow suit — or will they leave them behind for the next incumbent to retrain them? And as large sections of creative workforces deploy their digital personalities, intellectual property lawyers will have a field day separating the personal mannerisms that are part of the individual brand from the content that may be a corporate asset. 

There are some middle-of-the-road scenarios as well: WorkFusion was born in 2010, producing bots that helped automate the excruciatingly mundane of financial processes and was one of the first in what is now a red ocean of automation. But VC-backed and genAI-fuelled in the post-Covid era, WorkFusion progressed into a supplier of context-specific ‘AI digital workers.’ From transaction screening analysts to due diligence analysts and adverse media screening analysts, WorkFusion today has a growing corporate client base whose employees get augmented as and when needed and for as long as needed, often working with minimal supervision and indefinitely. While these hireable AI digital workers are neither ‘fused’ nor are avatars, their proliferation could raise some questions in a similar strain to that of digital-human fusion. 

Will genAI alienate CIOs? 

But by far the biggest consequence — laterally speaking — of the AAA journey is the potential for disintermediation of the CIO and the IT and tech offices in any enterprise. Digital twins are getting built and born in the enterprise without any CIO facilitation. Funding for this is minimal at first and likely to grow — outside of the IT budget. Analyst estimates of IT and tech spend outside of the CIO/CTO budgets range from 40-45-to 60%. In October 2023, forever fighting in the CIO’s corner, Gartner proclaimed that “45% of CIOs are driving co-ownership of digital leadership.” This is a sweetened narrative of the shift of IT dollars directly to users, business leaders and corporate functions, and it’s a move that genAI will exacerbate. 

The key prerequisite to make this a robust and sustainable trend is the ability to define an output: If you, as a business user, can succinctly define what output or outcome you need — as in Jesch’s case, represent him verbally and visually for a specific function — genAI can help, with the right prompt engineering, to deliver without elaborate technology pomp and ceremony…at least in the initial phases. In fact, this addresses the chronic inability of business to define desired outputs and outcomes that has kept IT projects from meeting business requirements. GenAI can ‘speak business’ without technical baggage. 

It is in this ability of genAI to become, in effect, an abstraction layer between the complex pipework and syntax of modern tech and the articulation of a simple business output, that a future of genAI lies. At least in this scenario, the democratization of technology will compel CIOs to attend more to the foundational tasks of redefining data architectures, dealing with the current data center resurgence and the realignment of many more software and hardware stacks to make that abstraction practical. With 2025 IT spend forecast to cross $6 trillion, there is much for both business and technology to plan for. 

Of course, not all genAI will necessarily touch business users and be in the prompt-and-click genre. Behind-the-screens systemic genAI and AI-to-AI engines, as well as AI that lives in the cloud, will be the lower-level detail that will drive ‘front end AI’ — and which will continue to remain a core technology domain. 

Many digital twins will replicate machines and processes, and many avatars will remain fabricated to fit a made-up profile of a standalone persona. But it is the augmented human — personified in the contemporary image of a Tumi backpacked manager — that will become the most interesting emblem of the genAI future.  

Picture this: Software repositories like Github and Stackoverflow have been mature enough for creatives to freely leverage and accelerate their work for many years — a kind of digital backpack for developers. And soon, as genAI-prompted code writes itself to go straight to Jira, Github Copilot’s licenced offering is raising eyebrows, to say nothing of the two years that a class action lawsuit has been in action. 

The questions these raise require collective enterprise and policy might. How will augmented and non-augmented employees work together? What management strategies will steer diverse digital-human teams? Compensation, incentivization and psychometrics, as well as legal and ethical frameworks, will be the issues to address — starting as soon as 2025! 

For over 30 years, Bhawani Shankar has worked across industries initially as a deputy CIO, then as an analyst and consultant — including eight years at Gartner — followed by business leadership roles with tier one technology services firms including HCL Tech, Cognizant and Wipro. Since 2022, he has been CEO at Cubix, an advisory firm that helps enterprises select and buy AAA products and services.