Using AI to Bypass Expert Bottlenecks in Strategic R&D

The Citrini experiment shows a shift in the corporate landscape: AI is lowering the barrier to entry for cross-disciplinary innovation so much that firms are starting to bypass traditional expert-led research in favor of "Gonzo" experimentation. This is not just about efficiency. It is about firms using AI to enter adjacent industries to validate their own investment theses. For leaders, the strategic advantage lies in using AI to perform work that was previously economically or operationally impossible, rather than just doing existing work faster. Those who wait for industry consensus or expert validation will be outpaced by competitors who use AI to generate their own proprietary ground truth.

The Illusion of Expert-Led Strategy

Traditional business strategy relies on a research-first model: gather experts, synthesize opinions, and then act. Citrini Research, a financial firm, found this model broken in the biotech sector. When they surveyed experts on AI potential in drug discovery, they received a polarized spectrum of answers, from "AI is useless" to "AI will cure everything in five years."

The system responds to ambiguity by stalling. Most firms would have stopped at the survey phase. Instead, Citrini used AI to bypass the expert bottleneck entirely. By treating the AI as an agentic partner rather than a search engine, they moved from passive analysis to active experimentation.

> "Our virtual biologists may have compressed a year's worth of work into a couple days, but biology works at the speed of life. DNA had to be synthesized, proteins produced and purified, toxins obtained in usable form, and appropriate controls established. Our digital invention was now set to create work for human scientists."
>
> -- Citrini Research

This reveals a systems-thinking insight: AI can collapse the dry lab portion of discovery, but it creates a new, rigid bottleneck in the wet lab physical world. The competitive advantage does not come from the AI speed alone. It comes from the willingness to accept the friction of physical-world validation that others are too risk-averse to touch.

The Inevitable Conflict of Vertical Integration

There is a growing debate about whether AI labs like Anthropic or OpenAI will eventually cannibalize their customers by moving into specialized fields like biology. The Citrini case study suggests the reality is more nuanced. AI labs are currently building tools to support existing players, but as the cost of entry for complex R&D drops, the distinction between a software company and a biotech company will blur.

> "I don't think this at all settles the question of whether at some point AI customers will find themselves competing with the AI labs, in fact that is going to be absolutely inevitable, but thinking it's going to happen all at once or across entire industries probably doesn't reflect the reality."
>
> -- Nathaniel Whitimore

The implication is that AI labs will not become biotech firms overnight. Instead, they will slowly internalize the value chain by providing the intelligence layer for increasingly specialized, high-stakes tasks. Companies that rely on AI labs for their secret sauce may find that their provider is also their future competitor.

Why Cures Won't Fix Public Perception

There is a belief in the tech industry that if AI can deliver a miracle, like curing a major disease, public sentiment will shift favorably. Systems thinking suggests this is a category error. Public perception is not driven by objective utility, but by narrative and trust.

If AI-driven drug discovery becomes the norm, it will simply be absorbed into the background of modern life. Just as we do not praise the miracle of modern logistics every time we order a package, we will likely view AI-assisted medicine as an expected utility. The headline strategy is a short-term distraction. The long-term reality is that AI must prove its value through consistent, boring, and reliable integration into existing workflows.

Key Action Items

• Audit your Expert Bottleneck: Identify areas where your team is stuck in analysis paralysis because of conflicting expert opinions. (Immediate)
• Run Gonzo Experiments: Allocate a small percentage of your R&D budget to execute high-speed, low-cost experiments that bypass traditional industry gatekeepers. (Next 3 months)
• Map Your Physical Bottlenecks: If AI speeds up your digital workflow, identify the physical or human-dependent speed of life constraints that will now become your primary limiting factor. (Next 6 months)
• Evaluate Vendor Dependency: Assess whether your core AI service providers are building tools that could eventually allow them to compete directly in your specific niche. (12-18 months)
• Shift from Efficiency to Capability: Stop asking "How can AI make this task faster?" and start asking "What work are we currently not doing because it is too expensive, that AI could now make viable?" (12-18 months)