Re-reading Geoffrey Moore’s Classic in India’s Age of AI

Geoffrey Moore, now 79, hasn’t formally updated his most influential idea, “the chasm,” for the AI era. Crossing the Chasm defined a generation of tech strategy, but as artificial intelligence has gone mainstream, Moore has chosen reflection over revision. In a Forbes interview, he acknowledged that today’s gap isn’t just about early adopters and laggards. It is about building trust and accountability in a world where technology moves faster than ever.

For founders in India, where AI is woven into daily life, from UPI payments to WhatsApp bots, the classic chasm doesn’t always show up where we expect. The old map helped us cross from skepticism to adoption. Now, we need to draw new lines to navigate from adoption to real alignment and value.So how did Moore’s framework sharpen the journey, and where must it bend in this new terrain?

Part I: The Origin, a Canyon Carved by Cost and Caution

When Crossing the Chasm landed on product-team desks in 1991, the very idea that a piece of software could reach “the mainstream” felt radical. Dial-up modems screamed, IT budgets crawled through annual committees, and CIOs lost sleep over whether a server upgrade would stall payroll. The risk was immense, and the path to widespread adoption was shrouded in doubt.

Into this uncertainty, Moore’s book offered a lifeline: a map of human behaviour hidden inside a simple bell curve. He borrowed Everett Rogers’ diffusion curve and relabelled its tribes to fit the quirks of B2B tech in those early years:

• Tech Enthusiasts (≈ 2.5%): tinkerers who will buy the beta just to break it.
• Early Adopters (≈ 13.5%): visionaries hunting strategic advantage, happy to sponsor a new category.
• The Chasm: the credibility gulf between promise and proof.
• Early Majority (≈ 34%): pragmatic buyers who wait for references, standards, and ROI calculators.
• Late Majority (≈ 34%): conservative buyers who adopt only when risk is near-zero.
• Laggards (≈ 16%): skeptics who switch only when the old system dies.

By insisting on psychographics over demographics, Moore gave founders a probabilistic lens. Market share wasn’t destiny, it was a game of staged evidence. But this chasm went beyond the psychological into a physical and financial reality, enforced by three factors:

• High-friction deployments: installing early PC-era software meant heavy capital expenditure, extensive training, and vendor lock-in. Switching costs made pragmatists justifiably cautious.
• Slow feedback loops: without cloud telemetry, it took quarters, not hours, to know if a product actually worked.
• Scarce social proof: analyst reports and conference keynotes acted as the primary trust brokers. Until they blessed a product, no one wanted to be the first to leap.

Put simply, the chasm was a rational, necessary pause for markets to verify claims in a world where evidence travelled slowly and the cost of being wrong was prohibitively high.

Part II: Redux, the Shifts (2004 and 2014)

The “chasm” wasn’t a static crevice. It evolved with every shift in technology and market structure. Moore revisited his framework, first in 2004 and then more sharply in 2014, to reflect how the adoption journey was changing under founders’ feet.

2004: Fresh Paint, Same Walls

The 2004 “Revised Edition” updated case studies, with Salesforce, Palm, and other early web-era names joining the classic cast. The technology was moving faster, with web-based apps starting to challenge install-heavy software, but the fundamentals of crossing the chasm stayed much the same. In India, these were the years of mass mobile adoption, but software “stickiness” and category-making still followed the original playbook: build trust slowly, leap when the herd moves.

2014: Cloud, Mobile, and the Shallowing Chasm

By 2014, Moore had to reckon with a digital world moving at cloud speed. His third edition (“Updated and Expanded”) did more than swap company names. He pointed to three shifts:

• Switching costs had melted: freemium, SaaS, and cloud-native products let users try (and abandon) solutions with minimal commitment.
• Churn replaced inertia: adoption was no longer the only hurdle. Retention, ongoing value, and rapid reference-building became existential.
• Consumerization: the adoption curve didn’t just describe CIOs. It also applied to everyday users, whose standards and switching behaviour could shape categories overnight.

India’s inflection: By the 2010s, India’s startup ecosystem was beginning to mirror these global shifts. Flipkart and Ola rewrote the rules for mainstream adoption, and millions sampled apps on cheap 3G data packs. Product-market fit became the raging buzzword. It could be tested, lost, and fought for over a year or two, not a founder’s lifetime. Experimentation cycles shrank, but true traction still required patience.

Part III: The Now, a Gradient Smoothed by Speed

Fast forward to January 2026, and that gap, or at least the version Moore drew, looks almost quaint. My neighbourhood aunty asks ChatGPT to rewrite her kitty-party invites. Freight brokers in Jaipur run half their back office through WhatsApp bots. The very idea of friction seems dated. NASSCOM’s 2024 AI Enterprise Adoption Index finds 70% of surveyed Indian enterprises now allocate over 20% of their IT budgets to emerging technologies such as AI.

Three structural shifts, particularly visible in India, have sanded that cliff into something closer to a speed-bump:

• Zero-friction distribution: browser-based SaaS, mobile app stores, and API-first tooling now ride India’s public digital rails, from Aadhaar-KYC to UPI’s 21.6 billion transactions in December 2025, letting users sample value in minutes.
• Ambient AI literacy: Salesforce’s 2025 generative AI snapshot shows 73% of Indian adults online already use generative AI tools, while Microsoft research reports 92% of knowledge workers in India leverage gen AI at work.
• Continuous trust signals: real-time reviews and open-source repos create rolling social proof. The buyer no longer waits for Gartner to bless a category.

When experimentation is free and evidence is instantaneous, the “show me ROI first” mindset of the early majority can appear on Day One. Adoption now feels more like a gentle gradient than a terrifying gorge.

So, if the gulf to initial adoption has narrowed, where did the danger go? It has migrated from adoption risk to alignment risk. A new chasm persists, less about buying and more about believing and benefiting:

• The Trust and Safety Gap: will this AI respect my data boundary, my compliance regime, my brand voice?
• The Integration Gap: can the new tool interoperate with the maze of legacy stacks that still run critical operations?
• The Outcome Gap: does the user possess the strategic clarity to turn raw capability into measurable advantage?

Part IV: The Way Forward, Stewardship in the Age of Agentic AI

Moore’s genius was never the specific percentages. It was the call to sequence evidence. In 2026, that call remains essential, while evidence now flows at gigabit speed and even skeptics carry smartphones.

But what does stewardship look like in the era of agentic AI, when technology is not just a tool, but a semi-autonomous agent shaping, nudging, and learning from its users? By “agentic,” I mean systems that can plan and act across tools with minimal supervision, not just answer prompts.

Continuous stewardship now means designing systems that adapt, audit, and align in real time. It means monitoring not just whether a customer is using your product, but how their agent is representing them, protecting their data, and acting on their behalf, sometimes making decisions at a speed and scale that outpace human review.

For Indian founders building in this terrain, defensibility will depend less on crossing the adoption gap (which is now shallow) and more on earning sustained trust as AI shifts from passive assistant to proactive collaborator. The new chasm is about agency, alignment, and accountability.

The challenge is no longer, “How do we cross it?”

The real question now: How do we remain worthy of trust, when our users, and their agents, can choose, adapt, and move on at the speed of a swipe?