Strategizing AI??It’s like solving a Rubik's Cube while it invents new colors.

"Exploring AI Futures with Bill Murray (not the actor)."

Will?Generative AI change business, the world's economies and societies more rapidly than the Internet? Sure, but at least the Internet let us?procrastinate with tiktok.

Will the?cost of AI-enabled software development plummet, savaging the business models of software engineering firms? Maybe, but at least AI will help me code my replacement.

Will emerging AI Agents (Introducing GPTs ( openai.com ) )?become how citizens and consumers get jobs done? And will we be able to find the kill switch if needed?

IF AI is so clever at automating everything, can you build a 'fearsome enjin' that cranks out AI solutions? And will it unionize before we can turn it off?

Amid the buzz, the critical question for firms is: where to go with the hundreds of?AI-enabled use cases in your firm? Think Small ,Start Big ,Test Slowly, Scale Leisurely?

No one wants to tell their board they missed the opportunity that AI offers organizations.

Even with change, certain patterns endure. This includes an approach shown on Wardley Maps – a concept from my DXC colleague, Simon Wardley . Earlier this year, we built a set of Generative AI maps with broad participation across the AI community. These depict the evolution of Predictive and Generative AI, while revealing three workable themes.What do you look at when change overtakes action?

We are seeing many?AI strategies come and go in real-time, in the news and through browsers. Many thought leaders have virtual trash bins heaped with articles promoting the 'AI next practice.'

The goal isn't just AI adoption but intelligent augmentation.

Another colour on your Rubik's Cube.

In fast-paced evolutionary change, some patterns stay the same - the stages of evolution – Genesis, Custom, Product and Utility. You see these on a?Wardley Map invented by my brilliant DXC colleague?Simon Wardley . Putting the two flavours of AI, Predictive and Generative on a Wardley Map shows these persistent patterns. For nine months, we have followed the evolution of AI using Mapping, and we found three persistent themes.

Three persistent themes in charting AI's evolution?

The Expansive Utilization of AI?

• Generative AI isn't additive - it's the main event.?Predictive AI animates sensing, operating, and analyzing. Generative AI almost trebles the range of Knowledge Tasks by enabling creating, designing, empathizing, conversing, and summarizing.
• We're experiencing a Cambrian Explosion of use cases. Automating these additional Knowledge Tasks means use cases exist across all functions of an organization, across all stages of the value chain, and across all participants of an ecosystem. AI now enables?fleets?of cyber-physical assets like cars, wind turbines and Manufacturing lines to tell us so much more about the real world, people and the unknown. It powers "data?flywheels"?to create new products, functions, and insights.??
• It will re-engineer software engineering. Bringing intelligent augmentations will reshape the data science process; creating synthetic data and documentation; designing code and developing architectures. It also analyses user needs, converses with users, and summarizes performance. As programming eases into conversations, the next wave of citizen developers will emerge. In the future, software development and costs may become so simple and cheap, the job market for these skills evaporates.?

The Mandatory Operationalization of AI?

• MLOPs evolves to IntelligenceOPs (IOps)?- Generative AI adds more complex, fast-evolving layers to the AI stack. Where we used MLOPs to automate the AI value chain, we now find intelligent augmentation enabling IOps – a more comprehensive set of capabilities to operationalize and help scale all types of AI. IOps gives companies the workflows to build, manage and govern a full spectrum of intelligent augmentation capabilities at scale. It evolves processes to harness unpredictable generative AI along with traditional AI.?
• "We'll need a bigger Boat"?- IOps expands the scope of every aspect of managing AI; Knowledge Tasks, Data, Models, monitoring and evaluation, deployment and governance.?
• Ethical AI evolves through stages. Technology in different stages of maturity requires different ethical and technical approaches. Every organization and community will grow to encompass an embedding of ethical requirements and governance - completed more "by feel" in the early stages, but more formalized in the more asset-driven later iterations.?

The Aggressive Industrialization of AI?

• AI resembles Matryoshka dolls?– Most people only interact on a Knowledge Task within the Applied Model, the Outside Doll; the other layers or dolls are contained within and need no interaction. However, the AI groups in an organization must grasp what to develop internally or procure externally, how to integrate and interface with these components seamlessly, and how to ensure their ongoing evolution and maintenance.?
• AI is Industrializing at speed –?this technology is singular in its rapid evolving from new technology experiments through Customized solutions through Product to Service. No other technology has industrialized in such a compressed horizon. The smart way to understand, guide and exploit this swift industrialization is to create a map of your AI projects to determine the levers you can pull.?
• AI is becoming the IT environment?- AI is 'appifying'; we now have user-friendly 'AI agents' to do tasks for us. AI is an Operating System; IOps ensure AI models are managed and served, similar to an operating system managing and serving applications. AI Factories become hubs where intelligence-driven technologies are developed and run - becoming the proving ground for e.g., training autonomous vehicles, robotics platforms, and large language models.???

Oh no, more colors on the Rubik's Cube?

There are some signals amongst the noise. While the timing is uncertain, developments and ambitions suggest the following:?

• Models will continue to leapfrog expectations, producing higher quality, more coherent outputs across modalities including text, images, video, and audio.?
• Models can combine multiple data types -- whether?text, images, or video -- and deliver more sophisticated results. This fusion will enable richer storytelling and more interactive experiences.?
• User interfaces and APIs will improve non-experts' access to generative models. AI will become apps and Agents to perform more complicated tasks on our behalf.??

AI will:?

• become more autonomous, self-teaching and self-experimenting.?
• power fleets of software-defined assets to avoid breakdowns, self-heal and evolve new functionality.?
• Enable non-expert access, making AI more autonomous and functional.?

• AI factories will produce intelligence-driven products such as trained models, smart interfaces, and basic AI apps.?

We posit the most realistic view of these extraordinary phenomena is to develop strategic options through the lens of Staged Evolution. Wardley Maps will allow you to predict what colors may be coming next and to invest appropriately to create and extract value.

The great eight: consider these specific actions?

1. Convene functional experts, data scientists, and AI-familiar ethical and legal experts to identify low-risk/high-reward priority areas for intelligently augmented applications.??
2. For each priority area, review your alliances and skill sets to decide on a fine-tuning vs training approach.?
3. Craft and communicate acceptable use policies for AI-powered solutions.?
4. Experiment with Generative AIs and learn how they answer, augment, and automate tasks.?
5. Identify and corral valuable data to establish competitive advantage with tuned models.?
6. Review your ecosystems for Generative AI capabilities to extend your products/services – develop partnerships and ecosystems.??
7. Grow your MLOPs to 'Intelligence Ops' to intelligently augment at scale, ethically.?
8. Define your Open vs Proprietary mix (likely with input from counsel.)??

Understanding AI’s evolution through Wardley Maps offers a strategic lens to anticipate the future landscape of AI – a necessary tool in a world where AI is not just a solution, but a constantly evolving challenge.?

=/=/=/=/=/=/

Discover more https://dxc.com/leadingedge

Embracing modernization: from technical debt to growth

Platform culture: a field guide to mastering the art of information flow

Engage with our researchers Bill Murray , Simon Wardley , Chris Daniel

?

?

?