TL;DR
At the recent Society of Decision Professionals (SDP) Annual Conference discussions highlighted a recurring challenge: frameworks alone do not improve decisions. Real impact depends on whether organizations turn data quality and structured decision-making into an organizational capability embedded across functions. Artificial intelligence (AI) will increasingly support this effort by helping structure evidence and unlock institutional knowledge—but long-term success still depends on sustained adoption across teams and therapeutic areas.
Intelligencia AI’s Marianna Esposito and Edoardo Madussi attended the recent Society of Decision Professionals (SDP) Annual Conference in Boston and came away with interesting insights and learning they are sharing in this blog.
Decision Quality Is Not Just a Framework
In pharmaceutical R&D, the concept of decision quality (DQ) has gained traction as companies look for ways to make portfolio and development decisions more structured and transparent. Frameworks for evaluating evidence, clarifying assumptions, and aligning stakeholders are now widely discussed.
However, a recurring theme at the conference was that the real challenge begins after the framework is introduced.
Many organizations successfully launch DQ initiatives, but their impact often remains limited to specific teams or pilot programs. Sustained value depends on embedding decision quality practices into everyday workflows across research, clinical development, regulatory strategy, and portfolio management.
This means, decision quality must evolve from a methodology into an organizational capability and broad adoption across functions is required for these tools to have long-term impact.
The Maturity Curve of Decision Quality
Another theme discussed during the conference was the maturity curve many organizations experience when implementing DQ approaches. Most initiatives move through three phases:
- Start: introducing frameworks, terminology, and governance structures
- Grow: expanding adoption across teams and therapeutic areas
- Maintain: sustaining relevance as portfolios, priorities, and leadership evolve
The first phase tends to receive the most attention, yet the later stages often prove more challenging.
As organizations evolve, new programs emerge and teams change. Without continuous reinforcement, decision quality practices can gradually lose visibility or become confined to niche groups. This dynamic explains why some companies struggle to scale decision approaches across portfolios or multiple therapeutic areas.
AI as an Enabler of Structured Decisions
No discussion about decision making in pharmaceutical development is complete without assessing the role of artificial intelligence (AI). Often AI is positioned as a replacement for expert judgment. However, this view is too simplistic, instead of replacing experts, AI is well-suited to help experts manage the growing complexity of evidence.
Potential contributions of AI include:
- structuring large bodies of scientific and clinical data
- synthesizing evidence from multiple sources
- retrieving relevant historical analyses
- supporting consistent evaluations across programs
These capabilities can help teams navigate increasingly complex decision environments more quickly and make them easier to scale across organizations while preserving the role of expert interpretation.
Institutional Memory: A Hidden Asset
One observation raised during conference discussions concerns the large amount of knowledge pharmaceutical organizations already possess: institutional memory.
Over decades of R&D activity, companies accumulate extensive documentation related to portfolio decisions, development strategies, and program outcomes. Yet this information often remains fragmented across reports, presentations, and internal databases.
At the same time, formal post-mortem analyses of failed programs are relatively rare, and lessons learned from past decisions are often not systematically reused. This institutional memory represents a largely untapped resource for improving decision consistency and avoiding repeated mistakes.
Expanding Opportunity Spaces Through Data
Another topic discussed at the conference involved the expanding range of data available for evaluating development opportunities.
Teams increasingly combine multiple types of evidence, including scientific and preclinical data, clinical trial results, and real-world data. When analyzed together, these data sources can highlight development paths that may not have been obvious within a single dataset.
Importantly, these approaches do not replace expert judgment. Instead, they expand the opportunity space considered during portfolio decisions while remaining closely aligned with expert insight.
The Real Challenge: Sustaining Adoption
Technology and analytical frameworks continue to evolve rapidly. Yet discussions at the conference suggested that the decisive factor may not be technology itself.
The real challenge lies in sustaining adoption over time.
Embedding decision quality across functions, maintaining institutional knowledge, and ensuring that analytical tools remain relevant as portfolios evolve may ultimately determine whether these approaches deliver lasting value and provide companies a real advantage in how they manage uncertainty in pharmaceutical development.
FAQ
What is decision quality (DQ) in pharmaceutical R&D?
Decision quality refers to structured approaches used to improve how evidence, assumptions, and trade-offs are evaluated when making development or portfolio decisions.
Why do many DQ initiatives struggle to scale?
Many programs begin as pilot initiatives or specialized teams. Sustained impact typically requires broader organizational adoption across functions, therapeutic areas, and decision processes.
How can AI support decision quality?
AI tools may help organize large bodies of scientific and clinical evidence, retrieve historical analyses, and structure complex evaluations. Most discussions emphasize AI as supporting expert judgment rather than replacing it.
For more tailored, data-rich insights, let’s talk.
