New: the AI in Pharma and Biotech Leadership Dashboard, Q4 2026
ResolveNext has launched the AI in Pharma and Biotech Leadership Dashboard for Q4 2026. This unique assessment shows progress on AI in terms of deployment, research, talent, and visibility at the global top 30 pharma companies by revenue, and the top 10% of US companies at each development stage.
ResolveNext data from 2,922 AI-related job postings show agentic AI design is the fourth most requested AI skill, behind automation, machine learning, and general AI.
The MHRA published a 2030 strategy that will apply AI to safety monitoring, and establish a new AI center and a trial-protocol checker.
The EMA is weighing the use of generative AI in drug manufacturing under a documented control strategy, but offers no method yet for measuring hallucinations.
Boehringer Ingelheim and Microsoft will test whether digital twins of patients can provide synthetic or hybrid trial control arms, starting in lung health.
ResolveNext Research · October 23, 2026
Agentic AI design ranks fourth among AI skills in biopharma job postings
ResolveNext analyzed 2,922 AI-related job postings from US pharma and biotech companies. It finds that 24% ask for skills or experience in agentic AI design. Automation and robotic process automation (RPA), at 40%, machine learning, at 33%, and general AI skills, at 31%, rank higher among AI-related skills.
Demand for AI skills varies by business area. Of the postings that require AI, 40% are for roles in research, clinical, regulatory, manufacturing, or quality teams. That's more than in digital and IT.
AI-related skills and experience are also becoming more important beyond dedicated technology roles. In research, 34% of the postings that require AI skills or experience are for scientists or traditionally less technical positions. In clinical, the number rises to 43%.
What it means for you. Agentic AI design now appears in 24% of AI-related postings, behind machine learning at 33%. Compare these figures with your own hiring plans, and decide whether each function in your organization is recruiting the AI skills it needs.
MHRA's 2030 strategy puts AI into safety monitoring and trial applications
The UK Medicines and Healthcare products Regulatory Agency (MHRA) published its 2030 strategy on October 14, 2026. The agency says it plans to use AI to identify potential medicines safety issues earlier, working with the Health Data Research Service on real-world safety data, and to make Yellow Card reporting simpler through a more conversational interface.
The MHRA will establish a Centre for AI and Digital Health to clarify expectations for developers of AI-driven medicines and medical devices, and will publish an AI medicines development action plan in 2027. An AI-enabled Clinical Trial Protocol Checker is planned to help sponsors identify common reasons for non-acceptance and to speed clinical trial applications. By 2030, the agency also aims to develop a model for regulating AI as a medical device.
What it means for you. If you run regulatory or clinical operations in the UK, track the agency's progress on the protocol checker and 2027 action plan, because they will help define the MHRA's stance toward the use of AI in trial applications and in drug development. Meanwhile, the AI center should offer a new place to look to for information and guidance.
EMA considers allowing generative AI in GMP with documented controls
The European Medicines Agency (EMA) published a report on October 2, 2026 from a multistakeholder workshop on Annex 22, the planned AI addition to the EU good manufacturing practice (GMP) guide. The workshop ran June 30 and July 1.
The current draft says generative AI, large language models (LLMs), and other probabilistic or adaptive models should not be used in critical GMP applications. Comments from the consultation supported allowing them, and the drafting group discussed doing so, provided they meet the annex’s requirements and are supported by a “fully documented, robust, risk-based control strategy.”
Concerns remain about what should be included in the control strategy. Regulators asked whether hallucination, fabrication, and sycophancy are ordinary statistical uncertainty or a different kind of failure, but the discussion did not lead to proposals for estimating confidence intervals or hallucination rates.
The report notes that in adaptive or generative AI, the inability to detect a failure may itself be a central risk factor. It also says that if a third-party model or application programming interface (API) cannot provide the outputs, rationale, logs, or evidence needed for control, the manufacturer cannot measure and justify the system adequately. The qualified or responsible persons remain accountable and cannot delegate this to AI, and human oversight cannot compensate for inadequate validation. The drafting group will review the input as it revises Annex 22.
What it means for you. If your manufacturing or quality teams want to use generative AI in GMP workflows, the annex may move from a ban toward a conditional yes, though the text is not final. Ask your suppliers for the logs, outputs, and change records described in the report, and confirm which of your qualified persons would be accountable.
Boehringer Ingelheim and Microsoft to test digital twins of patients in trials
Boehringer Ingelheim and Microsoft announced a strategic collaboration on October 22, 2026 to investigate the use of digital twins of patients in clinical development. The release describes the digital twins as computational representations of individual patients that use clinical and other data to forecast disease progression and how outcomes might change under different scenarios.
The partners will test whether synthetic or hybrid control arms, built from existing data, can replace or supplement a conventional control group. They aim to reduce the number of patients in control groups, collect clinical evidence more quickly, and improve trial design. The partners plan to start in lung health before expanding to other disease areas.
What it means for you. Synthetic and hybrid control arms could reduce the number of patients required for a trial. If you run clinical development, ask your biostatistics team where existing data could support a hybrid design, and what regulators would need to see.
Zealand Pharma has licensed Generate Biomedicines’ AI platform to design multispecific proteins for obesity, with $30 million upfront and nearly $1.2 billion in potential milestones.
Contract Pharma reports that GSK found binders for every target it tested in its own wet-lab evaluation of Chai’s zero-shot designs, and now has access to Chai’s platform.
Xaira reports that its X-Design AI produced seven confirmed binders from 92 selected candidates on a receptor target where two library screens had previously found nothing.
Ginkgo Datapoints and Apheris will build a 10,000-antibody dataset and train models on members’ proprietary data without exposing raw sequences, with initial access projected for early 2027.
ResolveNext · network-powered insight and advisory
