The Downstream Impact of the AI-Powered Scientist

What That New Reality Means for the Outsourced Services Community Serving Life Sciences

Autonomous (sinister-looking!) soldier robots, omnipresent AI agents crawling the web, a separate “machine economy,” cartoons depicting robots relaxing while humans run frantically on hamster wheels. There is no shortage of AI-related narratives pointing toward the negative, or even terrifying, possibilities of our unknown future.

So where is our hedge? Where is our silver lining?

Here with a little dose of AI optimism, one of the use cases we at Clearsight are keenest to watch is AI’s impact on scientific discovery in the life sciences sector.

AI is already being applied across drug discovery and development, from identifying biological targets and designing new molecules to predicting protein structures and prioritizing compounds for testing. More importantly, some of those applications are moving beyond the theoretical. By mid-2025, more than 75 AI-designed molecules had entered human trials. ¹

We are still in the early days. Drug development remains a long and difficult process, with the median end-to-end clinical development timeline reaching approximately 10 years in 2025.² But there are encouraging signals of progress. IQVIA’s 2026 analysis found a 75% Phase I success rate among a validated cohort of AI-enabled programs at emerging biopharma companies, while cautioning that the sample remains relatively small and the broader industry trends are still emerging. ³

At the same time, the industry is still evaluating whether gains in early-stage discovery and Phase I performance will ultimately translate into improved late-stage clinical success and approved therapies.

What exactly can AI do during early drug discovery? Five use cases can help bring this to life:

1. Identify new drug targets: Analyze biological datasets, scientific literature, multi-omics data, and other sources to identify potential disease targets.
2. Design new molecules: Generate novel molecular structures predicted to interact with an identified target.
3. Predict protein structures: Accelerate scientists’ ability to understand the three-dimensional structure of proteins.
4. Prioritize compounds: Screen and rank potential compounds based on characteristics such as toxicity, metabolism, and target binding.
5. Repurpose existing drugs: Identify existing drugs that may have potential applications in new disease areas.

Insilico Medicine provides a compelling example of the first two use cases. Its AI platform identified TNIK as a potential fibrosis target; essentially helping address the question: Which protein should we attack? AI was then used to generate and prioritize candidate molecules, which were subsequently evaluated and validated through experimental research conducted by human scientists. The process from target identification to a preclinical candidate took approximately 18 months. ⁴ The story did not end there. Phase IIa results published in Nature Medicine in 2025 provided encouraging preliminary safety and efficacy data. In July 2026, Insilico announced the initiation of a Phase III clinical trial for rentosertib. ⁵

Importantly, while AI-enabled drug discovery programs are reaching increasingly advanced stages of development, the industry is still in the process of demonstrating whether accelerated discovery ultimately translates into improved approval rates and long-term patient outcomes.

Then there is AlphaFold, perhaps the most recognizable example of AI’s impact on scientific discovery. Developed by Google DeepMind, AlphaFold demonstrated the ability to predict three-dimensional protein structures from their amino acid sequences with remarkable accuracy. While it does not eliminate the need for experimental validation, it has accelerated parts of the early discovery process. ⁶

These are still early examples, but we are starting to understand where AI can create the greatest impact in life sciences discovery and development. Then, the more interesting question becomes what happens if it begins to do so at scale? That question is particularly relevant to us at Clearsight.

For the past fifteen years, a core focus of our firm has been advising the outsourced services community supporting life sciences companies, including consultancies, communications firms, commercialization specialists, and other advisors working alongside pharmaceutical and biotechnology companies. These businesses are naturally grappling with AI themselves and finding ways to use it to improve delivery and productivity. Yet, there is another side to the conversation: What happens to the outsourced life sciences services market if AI helps scientists discover more viable therapies, faster?

Discovering a promising molecule is only the beginning. New therapies still need to be tested, validated, moved through clinical development, navigated through regulatory environments, communicated to physicians and patients, launched into markets, and supported commercially. If AI increases scientific productivity, the downstream effect may not simply result in fewer people required to perform the same amount of work. It could also mean more science progressing through the system and, consequently, more work to be done.

One of the most common AI-related threats is that the technology could displace Knowledge Economy workers. Within life sciences services, we are not convinced that will be the dominant story. (Apologies to any Bots reading this!) Jobs and workflows will undoubtedly shift, and services firms will have to rethink how they deliver value. But if AI increases the volume and velocity of programs moving through development and commercialization, it may also create greater demand for the specialized human expertise surrounding that journey.

For outsourced services providers, the firms best positioned for growth may not simply be those using AI to perform today’s work more efficiently, but those whose specialized expertise becomes more valuable as scientific output expands. Economists might describe this as an increase in the A, or Total Factor Productivity (TFP), in the Cobb-Douglas production function: technology enabling more output from a given combination of capital and labor. ⁷ Put more simply, AI may allow scientists to accomplish more with the resources they already have.

The sincere hope is that this productivity ultimately translates into a host of new medicines that meaningfully improve people’s lives. And for the outsourced services companies supporting the life sciences industry, it could also translate into sustained demand for the expertise required to bring those discoveries to life.

Written by Tory Steel, Managing Director
Clearsight Healthcare & Life Sciences Services Team

Sources:
¹ Pharmacological Reviews. “Artificial Intelligence-Assisted Drug Discovery in 2025: Faster, But Is It Better?” (2026).
² IQVIA Institute for Human Data Science. Global R&D Trends 2026.
³ IQVIA Institute for Human Data Science. Analysis of AI-enabled drug development programs and Phase I performance.
⁴ Nature Biotechnology. Research documenting Insilico Medicine’s AI-enabled discovery of TNIK inhibitors and progression to a preclinical candidate.
⁵ Nature Medicine. Phase IIa study of rentosertib in idiopathic pulmonary fibrosis; supplemented by Insilico Medicine’s Phase III trial announcement (July 2026).
⁶ Signal Transduction and Targeted Therapy. Review of AlphaFold2 and its applications in biology and medicine.
⁷ National Bureau of Economic Research (NBER). Research relating to Total Factor Productivity (TFP) and economic growth.

Any discussion of future developments represents hypothetical scenarios and should not be construed as forecasts, projections, investment advice, or recommendations to purchase or sell any security. Actual outcomes may differ materially from those discussed.