AI Is Not Ready to Be the Decision-Maker — And Pretending Otherwise Is the Real Risk
August 3, 2026
No Comments
Read More »
Close
The standard for investment allocation decisions supporting portfolios has tightened over the past decade. From 2015 through 2019, abundant private capital and receptive public markets allowed scientific innovation to carry disproportionate weight in financing and portfolio decisions. During 2020 and 2021, pandemic-driven liquidity, accelerated digital adoption, and investor demand for growth amplified this behavior. Financing expanded across biopharma company stages, and many early platforms were valued on the scale of their scientific possibility; the market rewarded optionality, pipeline breadth, and speed of company formation.
More recently, from 2022 to 2025, a reset occurred. Rising interest rates and a narrower financing window forced companies to reach decision-quality evidence with less runway, while R&D productivity pressure made long development cycles and execution variability harder to absorb. Investors and strategic buyers increasingly favored validated biology, visible clinical milestones, credible operating plans, and a differentiated path to commercialization.
Looking toward 2026-2028, capital may reopen selectively, but the investment standard is unlikely to return to the funding conditions of 2020-2021. Scientific probability remains foundational, but selective capital markets will likely require a more comprehensive investment case: speed to evidence, execution feasibility, regulatory clarity, competitive sustainability, and capital efficiency.
Despite this shift in investment behavior, many portfolio processes still rely on periodic reviews, manual synthesis, and governance cycles designed for a more forgiving market environment. This has resulted in decision latency: the lag between when evidence changes an asset’s economic attractiveness and when leadership changes investment allocation and / or portfolio positioning.
The issue is less about improved evaluation of assets and more about a system redesign of portfolio management and decision making, shortening the evidence-to-decision cycle so material signals are detected, interpreted, escalated, and converted into action before external pressure forces the decision.
Before further assessing the challenge of decision latency, we should first explore how portfolio decisions are typically made. Biopharma leaders already weigh a familiar set of dimensions when analyzing an asset: scientific probability, regulatory precedent, development velocity, operational burden, competitive positioning, and commercial scalability. However, what has changed is how comprehensively they need to be weighed together and the increased impact of capital efficiency into decisions.
Scimitar refers to this organizing lens as Probability-Adjusted Capital Productivity: not a new set of criteria, but a discipline for weighing the criteria stakeholders already use against the critical investment question of today’s market — does this asset still deserve the next wave of capital allocation given recent changes? This lens sits inside broader portfolio management operating systems to support evidence-driven decision-making relevant to current and near-term market priorities and investment behavior. How to apply that lens today is shown in Figure 1 and the subsequent bullet points:
Figure 1: Capital Productivity Framework
Applying the Lens Today
Figure 2 translates the dimensions above into a decision matrix; this output is designed to force collective assessment of asset economics, execution reality, and portfolio-level strategic value.
Figure 2: Capital Productivity Matrix
Knowing which asset deserves the next pool of capital is necessary, but it is not sufficient on its own. The other half of the equation is whether portfolio governance can act on that knowledge fast enough to matter, and that is a distinct, increasingly urgent capability. With the lens of capital productivity in mind, we can now explore how portfolio governance and operations have kept pace with the heightened rigor of standards from investors and the broader market. This would be a measurement of decision latency, which can be decomposed into four cumulative layers, each contributing to a critical part in the portfolio decision-making process:
A comprehensive assessment of decision latency with public data is challenging, as companies rarely disclose when a signal was first detected, interpreted, escalated, and acted upon. However, complementary data points and insights across development, operations, and capital allocation can demonstrate the increased challenge and magnitude of impact this has within biopharma.
The trigger behind these events changes constantly; several examples include financing scarcity, clinical failure, regulatory feedback, and M&A integration. This reflects how quickly market conditions themselves evolve, as shown in Figure 3.
Figure 3: Drivers of Portfolio Reallocation, 2023-2026 YTD
As the table above shows, the specific cause of portfolio reallocation rotates each period, reflecting how much market conditions themselves continue to evolve. This volatility sits on top of a portfolio landscape that has grown competitively intense: two-thirds of major pharma portfolios now target “herded” therapeutic areas, and the window to capture half a drug’s lifetime sales has compressed by 18 months over the last few decades.
This pattern is not unique to any single company, and a 2026 industry analysis of biopharma operating models explains why it matters. Many were designed on the assumption that strategic challenges arrive one at a time, in sequence, built around a once-a-year planning rhythm assumed to offer enough flexibility for whatever came next, with each challenge resolved before the following one appeared. That approach holds up when the operating environment changes gradually. However, it breaks down when pressures arrive concurrently and unpredictably, which is exactly the pattern Figure 3 demonstrates. The resulting gap is a structural mismatch between the flexibility the market now demands and the rigidity that traditional governance models still provide — not a failure of leadership courage or vision, but of operating-model design.
Decision latency is not a random operational quirk, but the direct, predictable symptom of a governance model still built for a slower, more sequential environment than the one biopharma now operates in.
In short, this combination of a genuinely unpredictable operating environment and a governance model still built for a more stable one is what makes this an enterprise-level concern. The stakes are not abstract: patent-cliff exposure alone runs into hundreds of billions of dollars industry-wide, and individual reprioritization events of the kind catalogued in Figure 3 typically disrupt tens to hundreds of millions of dollars in committed R&D spend and severance.
Each decision latency layer creates a different bottleneck, which means no single intervention (including AI) is sufficient on its own. For example, faster signal detection has limited value if interpretation remains contested; faster interpretation has limited value if governance waits for the next quarterly forum; and faster governance has limited value if the decision does not change capital, resources, or ownership.
Therefore, only a redesigned portfolio governance process and its supporting workflows can truly address the increasing challenge of achieving accelerated time-to-capital allocation. This would be anchored to three principles:
Process before technology
An AI capability layered onto a workflow with unresolved bottlenecks does not remove the bottleneck but only automates it. Process optimization and re-engineering of the underlying workflow (e.g., decision ownership, evidence triggers, capital-action mechanics) must be addressed on its own terms before considering the use and integration of AI.
None of this argues for speed at the expense of scientific judgment. There is a real difference between a decision that takes longer because the evidence genuinely needs more time to resolve, and a decision that takes longer because it is waiting for the next committee discussion.
Treat operational excellence as a valuation lever, not a support function
A redesigned workflow rarely stays contained to itself and can change the economics of downstream activities. Accelerated evidence-to-capital cycles can compound into more capital-efficient trials, more credible signals to acquirers and investors, and a portfolio that can take more shots on goal within the same capital pool. Some strategy-focused stakeholders may still consider this portfolio process redesign as an efficiency exercise rather than a driver of enterprise value. However, it’s both; this exercise is a corporate strategy priority to strengthen current and future capital decision-making.
Distinguish sustaining fixes from disruptive redesign
Much of what narrows decision latency today is sustaining in nature, such as automating evidence synthesis, standardizing scorecards, accelerating a stage-gate review that already exists; these activities should be funded and scaled quickly. However, the larger opportunity is disruptive: using AI to build a portfolio governance capability that could not have existed before the age of AI, replacing periodic review with continuous monitoring, manual escalation with model-informed alerts, and after-the-fact debate with pre-negotiated decision rights. This would be an entirely new way of making portfolio decisions that can potentially deliver a structural, competitive advantage. This distinction and how to operationalize it at the portfolio level is described in greater length in a related Scimitar paper called: The Portfolio Imperative: Converting AI Investment into Compounding Enterprise Value.
Biopharma’s capital-allocation challenges are not driven by science or market dynamics, but by governance design.
The highlighted probability-adjusted capital productivity gives executives the interpretation discipline to know which asset still deserves capital. However, any strong framework that is executed through a slow, burdensome process will still be late. Closing this gap means redesigning the workflow itself: fix the process before layering on technology, treat operational excellence as a valuation lever rather than a cost efficiency initiative, and fund sustaining fixes and disruptive capabilities to drive a transformative, impactful change.
If your organization is evaluating how to close the gap between disciplined capital allocation and the portfolio governance speed needed to act on it, that is precisely where this work begins.
For executive teams navigating the governance challenges outlined in this paper, Scimitar brings a combined corporate strategy and operational re-engineering approach to biopharma portfolio management. Executing the work through implementation and realized value, we help leaders identify where operational bottlenecks are eroding value and design and stand up the systems needed to close that gap.
[1] U.S. National Library of Medicine. ClinicalTrials.gov API v2 and study records. Trial phase, status, enrollment, geography, duration, and competitive-intensity inputs.
[2] U.S. Food and Drug Administration. Novel Drug Approvals for 2025; Drugs@FDA; oncology and rare-disease approval resources. Approval output and regulatory-precedent context.
[3] Sertkaya A, Beleche T, Jessup A, et al. Costs of Drug Development and Research and Development Intensity in the US, 2000-2018. JAMA Network Open. 2024;7(6):e2415445.
[4] Biotechnology Innovation Organization, Informa Pharma Intelligence, and QLS Advisors. Clinical Development Success Rates and Contributing Factors 2011-2020. February 2021.
[5] Getz KA, et al. New Estimates on the Cost of a Delay Day in Drug Development. Therapeutic Innovation & Regulatory Science. 2024. Used as context on the economic value of development time.
[6] Federal Reserve Board. H.15 Selected Interest Rates and FOMC historical materials. Macro context for changes in financing costs from 2015 through 2026.
[7] J.P. Morgan. Q1 2026 Biopharma Licensing and Venture Report. Evidence of selective capital-market reopening, later-stage concentration, licensing activity, and modality / therapeutic-area investment patterns.
[8] Public biotechnology and pharmaceutical company Forms 10-K and financing disclosures, 2022-2026. Evidence of pipeline prioritization, runway extension, development risk, competition, and partnership strategies.
[9] Yavuz AC, et al. On the Concepts, Methods, and Use of “Probability of Success” for Drug Development Decision-Making: A Scoping Review. Clinical Pharmacology & Therapeutics. 2025;117(4):967-977.
[10] U.S. Food and Drug Administration. Accelerated Approval Program; Rare Disease Evidence Principles; individualized-therapy draft guidance; and model-informed drug-development resources. Context for pathway clarity, confirmatory evidence, and regulatory flexibility.
[11] GeneOnline News. Biotech Layoffs in 2026: Trends, Triggers, and What’s at Stake. 2026. Monthly 2026 layoff data and company examples, including Passage Bio and enGene.
[12] Xtalks. Pharma and Biotech Layoffs 2026 Watch. 2026. Company-specific layoff and restructuring detail, including Theravance Biopharma and Passage Bio.
[13] Scimitar. The Portfolio Imperative: Converting AI Investment into Compounding Enterprise Value (Operating AI as a Portfolio in Biopharma, Part 2). Source for the sustaining/disruptive AI innovation archetype and its associated governance model.
[14] Fierce Biotech. Layoff Tracker 2023, 2024, and 2026 (annual editions). Company-level source for the 2023-2026 portfolio-reallocation trigger dataset underlying Figure 3.
[15] BioSpace. Waves of Employee Cuts Hit Biopharma in 2023; Layoffs Continued Across Biopharma in 2024. Company-level source for the 2023-2024 portion of the Figure 3 dataset.
[16] Drug Discovery and Development. Exploring the Forces Behind 2024 Biotech Layoffs: A Visual Journey; Mapping 2024 Biotech and Pharma Layoffs. Category-level source for the 2024 portion of the Figure 3 dataset.
[17] PharmExec. The Pharma Patent Cliff Explained: What It Is and How Companies Navigate It. 2026. Used as context on 2026-2030 loss-of-exclusivity exposure and strategic/M&A response.
[18] Drug Discovery News. Blockbuster Drugs Face a Massive Patent Cliff in 2026. Used for named at-risk revenue figures (BMS, Merck, Pfizer) informing the 2026-2028 forward view.
[19] DeepCeutix Strategic Briefings. $300 Billion in Pharma Revenue Loses Patent Protection by 2030. Used for 2026 M&A deal-value and deal-count context.
[20] Fougner C, et al. Herding in the drug development pipeline. Nature Reviews Drug Discovery. August 2023, Volume 22, Issue 8. Peer-reviewed; used for the finding that two-thirds of major pharma portfolios target “herded” therapeutic areas, up from 16% in 2000, and related life-cycle compression trends.
[21] PharmaVoice. Biopharma layoffs rise as drugmakers tighten belts and reorganize. August 18, 2025. Source for 2025 portion of the Figure 3 dataset.
[22] Xtalks. Pharma and Biotech Layoffs 2025. Source for 2025 portion of the Figure 3 dataset.
[23] Fierce Biotech. Layoff Tracker 2025. Source for 2025 portion of the Figure 3 dataset.
[24] Rapid Trials. Biopharma Layoffs in 2025: An Overview of Industry Challenges and Workforce Shifts. Source for 2025 portion of the Figure 3 dataset.
[25] GeneOnline News. Layoffs in Biotech and Pharma: A March 2025 Snapshot by Therapeutic and Technology Focus. Source for 2025 portion of the Figure 3 dataset.
[26] Tenthpin. Life Sciences Trends 2026: The Cross-Currents Challenge (extract from Life Sciences Trends 2026: The Era of the Smarter Operating Model). 2026. Used for the finding that traditional operating models assume annual planning cycles provide sufficient agility, and for the structural mismatch between environmental volatility and operating-model rigidity.
Greg Caldwell's work spans corporate, portfolio, and pipeline strategy, operational excellence, M&A and asset monetization, new product planning and launch, and enterprise AI strategy and implementation, covering the full arc of value creation across the biopharma enterprise. He has led pipeline prioritization and governance engagements, supported multi-billion dollar transactions, designed and deployed AI capabilities across R&D and commercial functions, and built the integrated launch and operational frameworks that translate strategy into measurable patient and business outcomes. Across each domain, his focus remains constant: equipping leadership teams with the strategic clarity, operational rigor, and capability infrastructure required to move faster and more decisively at the moments that matter most."
