The 70% Rule: Why Your AI Strategy Is Focused on the Wrong Things
The Safe Mistake That Is Costing You the Race
Every mid-market company in India is now doing something with AI. That much is no longer in question — the real question is what they are doing with it and whether it will matter.
The pattern is remarkably consistent. A leadership team reads about AI transformation, hires a consultant or appoints an internal task force, and produces a roadmap. That roadmap typically starts with the same three initiatives: an HR chatbot for employee queries, an IT helpdesk automation tool, and a document management system with search. Sometimes there is a procurement workflow thrown in.
These are not bad projects. They are safe projects. They reduce some friction, they generate visible deliverables, and — critically — they do not disturb anyone with real power in the organization. If the HR chatbot fails, the CHRO absorbs the hit. If it succeeds, leadership can point to an AI win without having touched the core business.
The BCG research is unambiguous about why this is a strategic mistake. Across industries and geographies, 70% of AI's measurable value comes from core business functions — product R&D, digital marketing, pricing and revenue optimization, and innovation processes. The support functions that most mid-market companies start with generate the remaining 30%. Companies that optimize the support layer first are not being cautious; they are spending the first two years of the AI era building in the wrong direction.
What the Data Actually Shows
BCG's research on AI maturity segments companies into three cohorts: future-built, scalers, and laggards. The differences between these groups are instructive not because they describe different levels of AI sophistication, but because they reveal fundamentally different strategic orientations.
Future-built companies — currently about 5% of the corporate universe — allocate 15% of their total AI budget specifically to agents. Agents here means autonomous AI systems that can execute multi-step tasks without human intervention: conducting competitive analysis, drafting market entry recommendations, building pricing scenarios. These companies are not just using AI to answer questions faster; they are using it to execute decision-making workflows. Thirty-three percent of future-built companies have meaningful agent deployment today, compared to 12% of scalers and essentially zero among laggards.
The trajectory matters as much as the current state. BCG projects that AI agents will represent 17% of total AI value today, rising to 29% by 2028. Companies that are not experimenting with agent-based AI now are not just behind; they are falling behind at an accelerating rate, because the compounding effect of learning — what works, what does not, how to integrate AI outputs into actual decisions — takes time that cannot be recovered later.
McKinsey's parallel research adds texture to the BCG findings. Marketing and sales saw the biggest surge in AI adoption between 2023 and 2025, effectively doubling their usage rates. This is not accidental — it reflects where the ROI signal was clearest. In marketing, AI reduces content production costs and improves targeting precision. In sales, it sharpens lead scoring and enables more personalized outreach at scale. Both are measurable in revenue outcomes within quarters, not years.
But the more important McKinsey finding is about rewiring versus layering. Among companies that report high AI performance, 55% have fundamentally rewired their core processes around AI capabilities. They did not bolt AI onto existing workflows — they redesigned the workflows themselves. Companies that layer AI on top of legacy processes capture a fraction of the available value. The interface changes; the underlying logic does not. That is not transformation; it is decoration.
The India Context Complicates the Picture
Indian enterprise AI priorities, as documented in the EY-CII joint research, show a deployment pattern that deserves scrutiny. Operations tops the list at 63%, customer service sits at 54%, and marketing at 33%. On the surface, this looks more core-business-focused than the HR chatbot pattern described above. But the details matter.
"Operations" in this context is dominated by workflow automation and ERP integration — reducing manual data entry, digitizing purchase orders, automating invoice matching. These are efficiency plays, not innovation plays. They reduce cost at the margin without touching competitive differentiation. A company that automates its operations layer is a leaner version of what it already was. That has value, but it is not the same as using AI to identify a new market segment before competitors do, or to dynamically price products in response to real-time demand signals.
The talent constraint shapes this distortion. India's EY-CII data shows that 59% of business leaders identify AI talent as their primary constraint, and 91% cite deployment speed as the most critical factor in AI success. When you are talent-constrained and under pressure to show results fast, you naturally gravitate toward the applications with the clearest technical path and the least organizational disruption. HR chatbots and IT helpdesk automation are well-understood problems with established playbooks. Using AI to build proprietary competitive intelligence or to automate pricing strategy requires much higher analytical capability and much more organizational courage.
This is the trap. The talent shortage pushes companies toward the safe applications. The safe applications generate visible but low-value wins. Those wins become the narrative of the company's AI journey, even as the competitors who took harder but higher-value bets pull ahead.
The 80% Efficiency Trap
McKinsey's data on AI objectives reveals another structural problem. Eighty percent of companies set efficiency as their primary AI objective. Among high performers — companies generating measurable financial returns from AI — the orientation is different: they focus on growth and innovation, not cost reduction.
This is not a coincidence. Efficiency is a finite game. You can automate away manual processes until you reach a floor, at which point the AI investment stops compounding. Growth is an infinite game — better competitive intelligence generates better market entry decisions, which generate revenue that funds more intelligence capability, which generates better decisions again. The compounding logic of growth-oriented AI investment is fundamentally different from the diminishing returns of efficiency-oriented AI investment.
For mid-market companies in India, the efficiency focus is partly rational — margins are often thin and operational improvement has real value. But the companies that will define the mid-market landscape over the next decade are not the ones that reduced overhead by 15%. They are the ones that used AI to identify underserved market segments before competitors, to price precisely enough to protect margin while gaining share, and to move from reactive competitive response to proactive competitive shaping.
Bain's research on the top five AI use cases globally — sales, software development, marketing, customer service, and onboarding — provides a useful anchor. Four of the five are customer-facing or revenue-generating. Only customer service overlaps meaningfully with the support-function focus of most mid-market AI roadmaps. The others require committing AI resources to the parts of the business where competitive advantage is actually built.
A Framework for Reorienting AI Investment
The practical question for a ₹100 crore to ₹500 crore company is not whether to shift AI investment toward core business — the data on why to do it is overwhelming. The question is how to do it in a resource-constrained environment without abandoning the efficiency work that is already underway.
The reorientation begins with a clear categorization of the current AI portfolio. Every initiative should be mapped to one of two buckets: efficiency (reducing cost in existing processes) or differentiation (improving competitive position in ways that generate revenue or margin). Most mid-market roadmaps, on honest assessment, are 80 to 90% efficiency and 10 to 20% differentiation. The target ratio for companies that want to compete effectively in five years is closer to the reverse.
The next step is identifying the specific differentiation opportunities where AI has the clearest leverage. For an industrial manufacturer, that might be AI-powered competitive pricing — monitoring competitor price movements across channels and adjusting in near-real-time rather than quarterly. For a consumer brand, it might be AI-driven demand sensing across tier-2 markets, where traditional research is expensive and slow. For a B2B services company, it might be AI-assisted account prioritization that identifies which clients are most likely to expand based on behavioral signals.
None of these require the organization to abandon its efficiency initiatives. They require leadership to consciously ring-fence a portion of AI investment — BCG's future-built benchmark of 15% dedicated to agents is a reasonable starting target — and protect it from the gravitational pull of safe, visible, low-value projects.
The Window Is Narrowing
The BCG, McKinsey, and Bain research collectively points to a window of differentiation that is available to companies right now but will not stay open indefinitely. The companies that are building AI-powered competitive intelligence, pricing engines, and go-to-market capabilities today are accruing learning and data advantages that compound. Their models get better. Their processes get faster. Their judgment about what AI can and cannot do gets sharper.
The companies spending those same months building HR chatbots will eventually run the same competitive intelligence and pricing initiatives — but starting later, with less accumulated learning, against competitors who have a head start.
The 70% rule is not a prediction about the future. It is a description of where value is being created right now, today, in companies that are winning with AI. The question every mid-market leadership team should be asking is whether their current AI roadmap reflects that distribution — or whether it reflects the organizational comfort of starting where it is easy.
The data has an answer. The harder question is whether leadership is willing to act on it.
LeanStrat helps mid-market companies redirect AI investment toward revenue-generating applications — competitive intelligence, pricing strategy, and market entry analysis. Get a free competitive scan at leanstrat.co/assessment.