A "10X" Change

For months after that Chicago flight, I carried the unease like a low-grade fever — not debilitating, not visible, but always there. I went to work. I ran my teams. I shipped features. But in the evenings, alone in my home office in the Marina, I did something I had not done since graduate school at Stanford: I went back to first principles.

I needed a framework. Not a feeling, not a hunch — a framework. Something that could tell me whether what I was sensing was real or whether I was simply a middle-aged CTO having an anxiety episode triggered by a clever chatbot.

I went back to Andy Grove.

Grove’s analytical model, which he adapted from Michael Porter, says that every business exists within a constellation of six forces: competitors, customers, suppliers, potential entrants, substitutes, and complementors. These forces are always in motion. Most of the time, they shift incrementally — a new competitor enters the market, a supplier raises prices, a customer’s needs evolve. You adjust. You adapt. You continue.

A Strategic Inflection Point is different. It occurs when one of these forces changes not by ten or twenty percent, but by a factor of ten. A 10X change. When that happens, the old rules stop working. Your strategy, your org structure, your product architecture — everything that was optimized for the previous equilibrium — becomes a liability. The things that made you successful become the things that kill you.

I had always understood this intellectually. Now I needed to understand it viscerally. So I sat down and mapped the six forces against what was happening in our industry. What I found kept me awake for three nights.

Substitutes changed by 10X.

Arcline’s procurement platform automated a workflow that had historically required highly trained humans — sourcing specialists, compliance analysts, contract reviewers. Our software made them faster. It organized their work. It enforced rules. But it still needed them. Every license we sold assumed a human on the other end, clicking through screens, making decisions, exercising judgment.

The substitute that had arrived was not another software platform that required slightly fewer humans. The substitute was not human at all.

A language model, properly prompted and fed the right context, could draft a procurement specification, compare vendor proposals against compliance requirements, flag anomalies in a contract, and generate a recommendation — all tasks that our platform required a trained human to perform. It could not do these tasks perfectly. But it could do them at a cost that was not ten percent lower than the human cost, or fifty percent lower. It was lower by orders of magnitude. We were talking about dollars per hour versus fractions of a cent per query.

When the substitute is not merely cheaper but essentially free at the margin, pricing strategy becomes irrelevant. You cannot compete on price with zero.

Potential entrants changed by 10X.

For thirteen years, I had taken comfort in the barriers to entry in enterprise procurement software. The domain was complex. The compliance requirements were labyrinthine. The integration work was brutal. Building a credible product required deep expertise and years of iteration. These barriers had protected us from competition more effectively than any patent or proprietary algorithm.

AI demolished those barriers in about eighteen months.

By mid-2024, I was tracking a startup — two founders, no employees — that had built a procurement workflow tool using a foundation model, a vector database, and a thin application layer on top. They had no domain expertise. They had never worked in supply chain. But they had fed the model thousands of pages of procurement documentation, compliance frameworks, and industry standards. And they had produced, in four months, a product that replicated roughly sixty percent of Arcline’s core functionality.

Sixty percent is not one hundred percent. But sixty percent at one-tenth the price, delivered in days instead of eighteen months of implementation, was more than enough to win the deals we were not paying attention to — the mid-market companies, the emerging buyers, the customers who had always wanted a tool like ours but could never justify the cost or the complexity.

This was not one startup. I found seven of them in our space by the end of 2024. Seven entrants that could not have existed two years earlier. The moat I had spent thirteen years digging had been crossed not by a more powerful competitor with a bigger army, but by a handful of people with laptops and API keys.

Complementors changed by 10X.

In enterprise software, your product is only as valuable as the ecosystem it plugs into — the ERP systems, the analytics tools, the communication platforms, the workflow engines. Arcline had spent years building integrations with this ecosystem. Those integrations were a competitive advantage. Every new connection we built made our product stickier and harder to replace.

AI redrew the complementor map overnight. When a foundation model can read any API documentation, generate integration code, and adapt to any data format, the value of hand-built integrations collapses. The new complementors were not the traditional enterprise software vendors we had partnered with for years. They were foundation model providers — Anthropic, OpenAI, Google — and the infrastructure platforms that hosted them. The center of gravity in the software ecosystem shifted, and our carefully constructed network of integrations went from being a moat to being technical debt.

The customer changed by 10X.

Not the customer’s identity. Their expectations.

In January 2024, I visited one of our largest customers — a global manufacturing conglomerate. The conversation started normally. They were happy with Arcline. Renewal was not in question. But toward the end of the meeting, their Chief Procurement Officer said something that stopped me mid-sentence.

“Vikram, your product is excellent. But my team has started using ChatGPT on the side. They paste our procurement specs into it and ask it to flag issues. It takes seconds. Your tool takes clicks, menus, and training. How long before your product works like that?”

She was not threatening to leave. She was not evaluating competitors. She was telling me, as a friend and long-time customer, that her team had tasted something and they were never going back. Once a procurement analyst has experienced getting an answer in thirty seconds by typing a question in natural language, they will never again tolerate navigating six screens and two dropdown menus to get the same answer from a legacy interface.

Customer expectations ratchet. They do not relax. And this ratchet had just clicked forward by a decade in a single year.

Here is what made this moment different from every other technology shift I had lived through. In the PC revolution, the substitute changed by 10X. In the cloud revolution, the delivery model changed by 10X. In the mobile revolution, the access point changed by 10X. Each of these was a single-axis transformation. One force moved dramatically, and the others adjusted around it.

With AI, four forces moved by 10X simultaneously.

Substitutes. Entrants. Complementors. Customers. All at once. All in the same eighteen-month window. This was not a Strategic Inflection Point. This was a Strategic Inflection Earthquake. The ground was shifting under my feet in multiple directions at the same time, and the frameworks I had used to navigate every previous technology transition — calmly evaluate, carefully plan, methodically execute — were designed for a world where you could afford to be calm and careful and methodical.

In a world where four forces move by 10X at once, calm is a luxury you cannot afford.

I did not know it yet, but Arcline was standing on a fault line. And the earthquake had already begun.