After the highs and lows of the nearly 70-year-old field of artificial intelligence (AI), it’s summer once again, with the rise of generative AI (GenAI). It’s been a scorching three years, with not a week seeming to pass without a significant technical breakthrough, eye-watering funding round, or widely optimistic claim about its potential to solve the world’s problems.
Yet, when we talk to board members and senior executives about what this all means for their businesses, the answers are often uncertain. Business leaders are broadly aware of both the opportunities and the risks, and many have been exposed to consultants and analysts versing them on the transformative power of GenAI – often with a stark warning: Do nothing and you will be left behind.
But what to do? Be bullish and invest in AI to proactively shape the future of the organization, or wait until the business benefits become clearer?
The nascent field of GenAI is surrounded by significant uncertainties that organizations need to navigate. These are not just technical, but economic, competitive, regulatory, ethical, and capability-related. This is not unusual in strategy formulation. Business leaders should not try to eliminate uncertainty (impossible), but instead make thoughtful decisions about where and how to invest. It is helpful to approach these strategic decisions with a structured framework. Our AI Navigator connects the arenas of business benefits (Where to play?) with the different types of AI models necessary for execution (How to capture value?).
There are three distinct areas of applications where GenAI can bring significant value to an organization: individual effectiveness, organizational effectiveness, and core innovation. The choice of how you build your AI portfolio will depend on the level of AI maturity of your company.
Individual effectiveness means using GenAI to increase productivity and augment the capabilities of your employees. The impact is at the task level, automating often repetitive, well-defined tasks, for example, summarizing information, generating options for marketing copy, or producing software code components. GenAI tools for employee effectiveness include powerful tools such as copilots, search engines based on large language models (LLMs), and smart assistants. The business benefits can be real if the gains are properly measured, and the effectiveness improvements are captured through increased productivity or redirected to different work tasks. Unilever’s customer service employees, for example, use GenAI tools regularly to draft responses to common customer inquiries. Partnering with Accenture, Unilever deployed more than 500 GenAI applications, including Accenture’s GenWizard platform, such as the AI-powered customer connectivity model that executes 13+ billion daily computations. In a pilot with Walmart Mexico, this resulted in product availability of up to 98%, and an approximately 12% sales uplift, driven by GenAI-enhanced forecasting and replenishment.
The impact of individual effectiveness initiatives is often localized and sometimes hard to scale beyond knowledge sharing between employees. As a rule, start here if your organization is at an early stage in its AI maturity.
Organizational effectiveness is a second-order impact beyond individual tasks, where a collection of tasks can be automated and embedded into the organization through redesigning workflows, processes, and even entire functions such as customer service. The business value can be substantial, but it requires a high level of organizational adaptation to capture the benefits. Redesigning technology-enabled processes across organizational silos, securely embedding them into existing enterprise systems, and, more critically, ensuring employee and team adoption of AI-driven workflows and new ways of working. Are the benefits scalable, for instance by replicating the new process across geography, or using the learnings to extend AI enablement to other processes or functions within the organization? Over time, such implementations can profoundly change the way an organization operates. Siemens, for instance, redesigned their entire industrial automation software development process by integrating AI coding assistants. Engineers complete programming tasks 30% faster, with junior developers showing even higher productivity. A broader study by Microsoft and Accenture found a 26% average task completion gain, especially in junior teams, without workforce cuts – reinforcing the Siemens finding. This is a potentially significant source of value for organizations, but it needs top-level commitment to execution. Here, we are in work redesign and change management territory, which are traditionally complex endeavors in large organizations.
Core innovation is an even higher order of GenAI enablement. The first two are about optimizing tasks and workflows within the boundaries of the existing operations. Core innovation is about changing the way the game is played: enabling fundamental change to a core part of value chains, to organizational structures, to value propositions, or even to business models. The business benefits can be a step change in the cost structure, revenue uplift, or GenAI-driven autonomous decision-making. Core innovation is a more complex and risky endeavor and, to be successful, needs to be undertaken as an enterprise-wide strategic initiative. Australia-based Canva, for instance, innovated away from being simply a graphic design platform to offer its Magic Studio, a GenAI-powered creative suite, in 2023. The business model shifted from template-based design tools to GenAI-assisted content creation and, in the process, expanded their market reach to millions of non-designer users, with significant increases in content creation volume. Here we are in the high-risk and potentially high-reward territory best suited for disruptors and for organizations with a high level of AI maturity.
Are these benefit areas mutually exclusive? No. Over time, executives can build a portfolio of AI initiatives spanning multiple areas of benefit. However, organizations go through a learning curve in terms of AI maturity. So, walk before you run!
“General-purpose, specialized, and pioneering AI models are the technology-enabled routes to value capture.”
There is a plethora of GenAI models on the market today that cover a wide array of applications. Ranging from well-known text and content creation tools (e.g., OpenAI’s ChatGPT, Anthropic’s Claude, or Google’s Gemini), to image and video creation tools (e.g., DALL·E 3, Midjourney, or Synthesia), to code assistance (e.g., GitHub Copilot, Amazon’s CodeWhisperer, or DeepSeek Coder), to audio and music generation (e.g., ElevenLabs, Suno, or Udio), and many others covering areas such as research, customer service, and data analytics, to name a few. The supply landscape is huge, and it’s easy to get lost. To navigate this landscape successfully, think in terms of three broad categories of AI model applications:
General-purpose AI models are off-the-shelf AI models, such as many of the examples mentioned above. They are designed to handle well-defined tasks with minimal customization beyond basic settings or prompting. General-purpose models can generate immediate value in areas such as document classification, customer support chatbots, or basic data processing. These models are widely available at reasonable operating costs, easy to implement, without the need for deep integration into a company’s unique business processes or systems. General-purpose models are often the backbone for targeting benefits at individual and organizational effectiveness levels.
Specialized AI models are specifically trained and/or fine-tuned for industry verticals, functions, or business domains, making them more powerful and relevant for narrower applications such as regulatory compliance in finance, medical diagnostics in healthcare, or supply chain optimization. Unlike general-purpose models, they require domain expertise and contextual adaptation to be effective, so implementation costs are higher. Specialized AI models bridge the gap between automation and business transformation at the organizational effectiveness level. They provide deeper insights and process improvements tailored to an organization’s needs and specificities. For example, OpenAI’s Harvey is a specialized model for the legal profession. It automates tasks like drafting, research, and contract analysis to streamline and enhance the efficiency of legal workflows.
Pioneering AI models are typically custom-built and/or highly experimental, often leveraging advanced AI techniques like reinforcement learning, generative design, reasoning-based AI, or agentic applications, which are designed to autonomously make decisions and act without human intervention. These are complex and costly to implement, with significant investment required in R&D, data infrastructure, and talent. Pioneering models target groundbreaking applications with novel solutions for domain-specific complex problems, such as drug discovery or autonomous decision-making in financial investment. Pioneering AI is the foundation for companies aiming to push the boundaries of competitive advantage by creating new ways to operate or business models. For example, Ocado, the online UK-based grocer, developed its Kinetic Storage, a pioneering AI model for the group’s automated warehouses, using generative optimization algorithms to dynamically reconfigure the storage grid based on predicted order patterns. The system has reduced retrieval times by over 30% during peak periods. The biotech company Moderna is building a pioneering AI model to design novel mRNA sequences tailored to specific therapeutic needs. These AI-powered tools are engineered to optimize key properties such as protein expression, molecular stability, and immune system activation, substantially accelerating the development of treatments for cancer and rare diseases. Notably, Moderna has partnered with IBM to apply advanced models like MoLFormer, aiming to improve the safety and efficacy of mRNA-based medicines through AI-driven design. Early results show promise in protein stability optimization, which could shorten development timelines by months.
As outlined above, individual effectiveness, organizational effectiveness, and core innovation are key areas for senior executives to focus on the value they want to extract from AI implementations in their organizations. General-purpose, specialized, and pioneering AI models are the technology-enabled routes to value capture. Both are important frameworks for executives to think through their strategy for AI. But, as with any strategy, the critical next step is to make choices by combining the right AI model with the right business benefit arena.
By taking a holistic perspective of the business objectives for AI, together with the technology-enabled models that can help realize these objectives, executives can frame the strategic choices that best suit the impact they are seeking for their organizations. The AI Navigator framework maps business value arenas with the types of AI model implementation, yielding seven choice archetypes that span quick wins to bold bets.
At the entry level, (1) Efficiency boost represents the intersection of individual effectiveness with general-purpose AI. In this model, humans oversee AI-assisted automation, with the primary business objective being the automation of repetitive, well-defined tasks to enhance personal productivity. These solutions rely on pre-trained models for document processing, chatbots, or email filtering, ideal for businesses seeking fast, low-integration wins. The benefits include reduced workload, fewer errors, and freeing up time for higher-value work. A typical use case is in insurance, where claims processors use AI to sort and categorize documents, allowing staff to focus on more complex cases.
When deeper domain knowledge is needed, (2) Smart augmentation pairs individual effectiveness with specialized AI. Here, humans collaborate closely with AI, aiming to enhance expert performance without redesigning workflows. Tools like AI copilots in legal or healthcare settings exemplify this model. For example, radiologists use AI to highlight abnormalities in medical images, improving diagnostic accuracy while retaining human judgment. Benefits include improved decision speed and reduced cognitive load – with AI amplifying, rather than replacing, expert insight.
Scaling up to organizational workflows, (3) Process accelerator sits at the intersection of organizational effectiveness and general-purpose AI. In this approach, humans supervise AI-driven processes to automate routine workflows across departments, often through tools like RPA and intelligent document processing. This strategy works well for firms aiming to streamline without deep structural change. For instance, a telecom provider automated ticket routing and account updates, eliminating bottlenecks and improving internal response times. Business benefits include enhanced throughput, reduced manual effort, and more consistent execution.
More mature organizations might pursue (4) AI-optimized operations, combining organizational effectiveness with specialized AI. Here, humans guide AI-driven decision-making across silos and functions. The objective is to redesign processes using AI that’s deeply integrated into operations, such as supply chain optimization or regulatory compliance. For instance, in the retail industry, a global apparel brand uses AI to align inventory, pricing, and demand forecasting across regions. This model breaks down silos, improves allocation, and enables real-time decisions, ultimately driving scalable efficiency gains.
For businesses aiming at a deeper transformation, (5) Operations AI step change emerges when organizational effectiveness meets pioneering AI. In this mode, humans oversee AI-led innovation, and the goal is to fundamentally reinvent how the organization operates. Techniques include AI-led forecasting, autonomous planning, and R&D support. Consider an automotive manufacturer using AI to predict market trends, simulate designs, and optimize factory layouts. This strategy improves agility, accelerates innovation cycles, and builds long-term advantage by reimagining how the enterprise functions.
For organizations aiming to accelerate their innovation cycles, (6) AI-powered innovation blends core innovation with specialized AI. Humans set the exploration boundaries, and the objective is to develop new business models or offerings without a full operational overhaul. Companies apply AI in drug discovery, product development, or design automation. In consumer electronics, for example, one firm analyzes customer feedback using AI to generate and test new product concepts, reducing R&D time, accelerating product cycles, and enabling responsive design with lower overhead.
Finally, for disruptors, or organizations that want to extend the boundaries of innovation in their industry, (7) Moonshot AI marks the boldest archetype, combining core innovation with pioneering AI models. In this approach, AI often operates with minimal human intervention, driving frontier innovation with agentic AI or autonomous AI applications. The aim is to create new value propositions or markets, often through autonomous discovery or decision-making. In biotech, for instance, startups are using generative AI to design novel proteins, developing treatments unimaginable with traditional R&D: the benefits include redefined markets, new competitive moats, and the transformation of entire industries.
Business leaders must focus their attention on where AI can meaningfully advance their strategic ambitions, given their current maturity, risk appetite, and investment horizon. This starts with a strategic choice about which outcomes to pursue; only then should you consider which technologies to adopt.
The AI Navigator doesn’t propose a single path forward – rather, it reveals the full spectrum of what’s possible. It helps shift the conversation from vague ambition to focused execution: Where can AI accelerate your current goals? Where might it reshape your longer-term operations? And where could it eventually reinvent your business model entirely? The AI Navigator framework invites reflection, alignment, and dialogue among executive teams. Used well, it can support sharper decision-making and help organizations craft a strategy for AI that is as ambitious as it is grounded.
Your strategy for AI will be iterative
Context-aware artificial intelligence is redefining the future of business. Amit Joshi shows how leaders can navigate through this radical shift to create new value propositions using a three-part framework. ...
One year after launch, AI risk tracker underlines worrying implications of accelerating AI adoption....
Fears that AI will dehumanize work overlook its potential to make organizations more human. Here’s how organizations can harness the power of human-centered AI....
OpenAI’s research shows AI hallucination is unavoidable. Leaders must adapt strategies, using limits as opportunities for innovation and resilience....
Explore first person business intelligence from top minds curated for a global executive audience