The core tasks of strategy work—collecting and analyzing data, generating and evaluating options, and synthesizing these insights in presentations and memos—are all information-intensive, making them a natural target for AI automation.
Yet the true impact of AI for the corporate strategy function will reside beyond task automation. As AI is adopted across the firm, new possibilities for strategy development will emerge, shaping how insights are gathered, where strategy is set, and how resources are allocated.
Harnessing AI to alter the system of strategy development itself will be the true differentiator.
The State of AI in the Corporate Strategy Function
BCG Henderson Institute analysis shows that more than 80% of the tasks strategists commonly engage in face either high or medium exposure to AI automation and augmentation. (See “The Strategy Function Is Highly Exposed to AI Disruption.”)
The Strategy Function Is Highly Exposed to AI Disruption
While mandates vary, most strategy teams act primarily as either a portfolio manager (40%) or a CEO office (30%), shaping corporate direction, guiding capital allocation, and supporting major strategic decisions. Other teams function as strategy orchestrators or internal strategy service providers.
Core activities span corporate and portfolio strategy, strategic planning, M&A and partnerships, capital allocation, performance monitoring, and executive decision support. The work is analytically intensive, process-driven, and largely episodic, structured around annual planning cycles and periodic reviews.
It is this concentration of information-heavy, model-based, and decision-support tasks that makes the function particularly exposed to AI disruption.
To identify which activities within the strategy function face disruption by AI, we analyzed task-level exposure data from O*NET, the US government’s occupational database, which assigns an “AI exposure rating” to a set of 41 work activities. We mapped these activities to tasks commonly executed within the strategy functions, which we compiled from several past BCG surveys of strategy leaders.
We found that strategy functions engage in 15 work activities tracked by O*NET and that more than 80% of these activities fall into the “high” or “medium” AI exposure categories.
- High-exposure tasks include getting, processing, and analyzing information, making decisions and solving problems, and working with computers, activities that constitute the bulk of traditional strategy work.
- Medium-exposure tasks include developing objectives and strategies, organizing and planning work, and various forms of communication and coordination.
- Low-exposure tasks cluster around the social side of strategy: organization development, fostering cross-business collaboration, and post-merger integration.
Many strategists are already using off-the-shelf tools, like ChatGPT, to generate competitor reports, brainstorm strategic initiatives, or outline compelling stories supporting the next five-year plan. Beyond this, powerful bespoke solutions are being developed to tackle specific strategy tasks:
- Some emerging competitive intelligence platforms can continuously scan more than 200,000 sources, notifying strategists of announcements from peers and policymakers.
- AI agents can develop financial models that outperform those of many first-year analysts.
- Emerging decision-intelligence tools can generate scenarios, simulating how different choices play out against real-world constraints.
The vendor landscape for strategy work is growing at a high pace: the first half of 2025 saw more strategy-related AI tools launched than the last two years combined.
This potential is not yet being systematically harnessed by most strategy functions. A recent BCG survey found that AI tools and agents have delivered consistent, positive impact for strategy leaders only in market intelligence and research. By contrast, more judgment-intensive, high-stakes activities related to M&A, partnerships, or portfolio management have not seen material improvements. Yet, our survey also found that strategy executives anticipate meaningful benefits in productivity, quality, and speed or outright automation across all their activities within the coming two years.
AI Will Drive Systemic Changes to Strategy Development
For the strategy function, the AI-first model will mean:
Supercharged Information Gathering. Successful strategy making depends on the broader organization for input. Yet, on average, roughly 60% of the strategy team is located in a centralized function, rather than distributed across the organization. As a result, insights often remain trapped locally, and strategies may be developed with missing or incomplete context. Exacerbating this issue is the fact that data may be stored in different formats or languages.
In the future, AI agents distributed across the firm could be used to gather insights across divisions, functions, and geographies to surface valuable insights. Generative AI can process unstructured data and thus act as a universal translator—synthesizing insights across fragmented and incompatible formats. AI agents may even summarize tacit knowledge that local teams observed but never formalized.
Over time, as agents in AI-first firms begin executing more workflows end to end, data could automatically be shared with the strategy function.
Decentralized Strategy Making. Strategy has historically been set centrally, as planning and analysis required specialized skills and dedicated capacity.
In the future, AI could equip teams across the organization to do the same work decentrally. When everyone has access to powerful tools that can support analysis, planning, simulations, and the communication of new options, strategy work becomes possible wherever the action happens—whether at the business unit or local level. Sanofi’s new tool called plai offers an early example of this potential shift. It enables more than 15,000 employees to run strategic analyses and what-if scenarios, allowing decision making to move closer to the frontline and leverage local experience and expertise.
As coordination costs decline, the fundamentals of strategic decision-making shift from largely hierarchical to bidirectional. Central strategy teams could devise a firm-level strategy by synthesizing emerging insights from across the organization. Meanwhile, decentralized decision makers (or AI agents themselves) could adjust course directly within agreed-upon guardrails.
Always-On Strategy. Today, strategy development and planning happens at regular intervals, often annually. This cadence is sensible because information gathering necessary for strategy development is time consuming and disruptive to day-to-day workflows in the line functions. However, by the time strategy has been formulated centrally, on-the-ground realities may already have shifted.
GenAI tools have the potential to remove this time lag, enabling teams to more easily maintain a continuously updated view of the strategic context. For example, tools like SignalAI monitor competitor filings, market news, and social media around the clock, enabling rapid decision making. Another example is L’Oréal’s TrendSpotter, a tool that continuously analyzes social data from more than 3,500 online sources to predict beauty trends six to 18 months before they become mainstream. This foresight allows L’Oréal to launch new products far sooner than traditional planning cycles would allow.
Agentic AI takes this even further by adding a layer of proactive intelligence. Rather than passively collecting information, agents can actively investigate what data inflows mean for strategic priorities and trigger followups or analyses. When incoming signals suggest a potential threat or opportunity, agents could autonomously pull relevant data from the firm’s knowledge stores and flag findings that require human input.
Importantly, this doesn’t mean that corporations or business units should overhaul their strategies on a weekly basis, confusing employees, customers, and investors alike. Rather, new real-time data-gathering tools could automate smaller refinements like adjusting prices, inventory levels, or production schedules. Meanwhile, more fundamental strategic decisions—like exiting risky verticals or acquiring a competitor—could be made faster than before and with more confidence as accumulating evidence validates the move.
Dynamic Resource Allocation. Strategic planning cycles decide the allocation of key resources including capital, budgets, and workforce. Once committed, firms face high friction costs for reallocation due to long approval chains and implications for corporate politics. As a result, resource allocation in corporations remains relatively infrequent.
AI automates much of the routine work that made frequent allocation impractical: monitoring existing performance and allocations, pulling and analyzing new data, and even suggesting shifts. This makes it possible for certain decisions that were previously locked in during periodic strategy planning cycles to become more operational. For example, imagine that always-on customer sentiment analysis detects a 15% spike in churn risk in a business unit. Automated tools could enable leaders to reallocate part of the sales force from new customer acquisition to retention within days, rather than waiting for the next quarterly review.
BCG Henderson Institute: Discover new thinking shaping the business landscape
How the New Strategy Function Can Add Value
In this new system of strategy development, new ways for the strategy function to add value emerge and will require different strategic capabilities.
Identifying the important questions and insights. As always-on and decentralized strategy making take root, a high volume of strategic insights can be generated continuously throughout the firm. As computation power becomes the new constrained strategic resource, strategists will need to decide which questions merit deeper analysis—an approach already used by Shell’s scenarios team, which deploys large models of the global energy system to substantiate long-term decisions.
But, of course, selecting the right questions is just the start. As analyses, insights, and strategic options roll in, judgment calls must be made. This ranges from sense-checking analyses to evaluating trade-offs between options. Many strategic choices lack objectively correct answers and involve navigating ambiguity, ethical implications, and trade-offs across stakeholders. This requires human accountability and judgment.
While fully outsourcing decisions to machine intelligence might be appropriate for some small-scale decisions, like operational resource allocation, the decisions fundamental to competitive advantage must remain human-led if they are to be differentiated. A new role for the strategy function could be designing decision-making systems that clearly define these boundaries, while ensuring decisions are made at the necessary pace and in a way that maintains strategic coherence and ownership of outcomes.
Given these shifts, we believe that core strategists will in many ways resemble today’s strategists: framing critical questions, evaluating options, and remaining accountable for high-stakes calls that define the firm’s direction. But the cadence and scope of their work will shift. They will curate portfolios of continuously flowing hypotheses and insights, shape how compute is deployed, and maintain coherence as strategy making becomes more decentralized.
Driving human-machine change. As always-on, dynamic, and decentralized AI systems generate a stream of opportunities and recommendations, firms will face a new challenge: a traffic jam of good ideas. Organizations, and their people, have limited capacity to absorb and implement change, creating a greater need for prioritization and sequencing.
Overcoming these constraints requires human involvement at two levels. Strategists must secure leadership buy-in on sequencing and make hard calls about transformation speed and its impact on people—work demanding both relationship capital and empathy. At the operational level, people must be brought along: for example, new workflows should be codesigned, building trust through involvement.
A new type of social strategist will be needed to coordinate this human side of the AI-first transformation. In many ways, the work of these social strategists will be familiar: driving adoption, securing buy-in, and overseeing strategic initiatives. Still, their focus will pivot toward overseeing the firm’s AI-first transformation: triaging opportunities, sequencing them in a clear narrative, and bringing people along to the new ways of working. Ultimately, they will be responsible for ensuring the human element moves at the pace the AI-first vision demands.
To complement this, specialized “technical strategists” will be needed to design and orchestrate the systems that make the AI-first strategy function possible. Most critically, they architect the autonomous decision-making systems by setting escalation criteria and governance frameworks. Working with IT, they might codesign agentic strategy workflows and data pipelines.
An AI-first transformation is not merely an exercise in enhancing efficiency: it will reshape how decisions are made, who makes them, and how processes and governance are designed across the firm. These are not primarily technological choices, but strategic ones. The strategy function is uniquely positioned to orchestrate where and how AI tools and agents are deployed within this shift, given its enterprise-wide vantage point. Chief strategy officers should therefore look beyond their own function’s adoption of AI and take on a broader role in shaping their firm’s AI-first journey—complementing technological upgrades with strategic leadership.
