01 Summary
A two-day strategic workshop translating an ambitious 2026-2029 platform vision into component level opportunities, agentic system themes, human-in-the-loop checkpoints and actionable roadmap inputs.
Lead: Product + AI Strategy
Product, Science, Engineering + Leadership
2 day offsite and synthesis deck
DMTL future state + agentic themes
I led a two-day offsite to help a cross-functional group make sense of how our sub-platform component should respond to the organisation’s wider 2026–2029 Platform vision. The aim was to create clarity, alignment and momentum. The overarching platform vision was ambitious, but each team needed to understand what it meant for their own area of ownership. My role was to design and facilitate a structured space where stakeholders could step back from day-to-day delivery, explore the implications of the future platform direction, and define a shared response that felt both strategically aligned and practically actionable. Over the two days, I helped the group translate broad platform ambition into component-level opportunities, priorities and downstream workstreams.
Challenge
The 2026–2029 platform vision created an ambitious direction, but component teams needed to interpret what it meant for their own users, workflows and future product direction.
The 2026–2029 platform vision created an ambitious direction, but component teams needed to interpret what it meant for their own users, workflows and future product direction.
Key questions
- Where could agentic capabilities genuinely support scientific workflows?
- What does the wider platform vision mean for this component?
- How do we avoid fragmented team-level experimentation?
System strategy
I translated the workshop outputs into a synthesis deck covering future-state workflows, automation opportunities, HITL checkpoints and roadmap themes.
I translated the workshop outputs into a synthesis deck covering future-state workflows, automation opportunities, HITL checkpoints and roadmap themes.
Strategy outputs:
- Updated 2029 DMTL future-state workflow
- Agent-supported capability themes
- Human-in-the-loop checkpoints
- Risks, dependencies and assumptions
=9
Workshop
I designed and facilitated a two-day offsite to align Product, Science, Engineering and Leadership around a shared response to the platform vision.
I designed and facilitated a two-day offsite to align Product, Science, Engineering and Leadership around a shared response to the platform vision.
Activities:
- Hopes and headwinds
- Knowledge sharing across component teams
- 2029 DMTL vision critique
- Component-level critique and playback
Impact
As this work sits at the strategic discovery and direction-setting stage, final product impact is still emerging, however early impact was:
As this work sits at the strategic discovery and direction-setting stage, final product impact is still emerging, however early impact was:
Immediate impact:
- Created shared strategic clarity
- Aligned teams around future direction
- Turned vision into actionable workstreams
- Framed agentic opportunity areas
- Reduced ambiguity around next steps
- Moved success metrics to sub-agent le
02 Context
A leading Tech Bio, founded in 2010, made a clear and ambitious bet: that the rapid progress seen in AI over the past decade would continue into the next, and that they are uniquely positioned to translate that momentum into meaningful impact for Drug discovery. In Q1 2026, they released their three year platform vision, with Autonomous Science as their North Star.
At the heart of the vision is the belief that drug discovery and development can become more adaptive, connected and self-improving. Their aim was to build an agentic system that could support end to end discovery by analysing data, making informed decisions, and designing the next best steps across the pipeline.
At the heart of the vision is the belief that drug discovery and development can become more adaptive, connected and self-improving. Their aim was to build an agentic system that could support end to end discovery by analysing data, making informed decisions, and designing the next best steps across the pipeline.
What made the vision powerful was its focus on grounding AI systems in patient data and strengthening the feedback loop between computational insight and experimental validation. In practical terms, this means helping teams move faster, make better-quality decisions, and use resources more effectively across the discovery and development process.
Ultimately, the goal was to create a platform consisting of a network of reasoning agents that helps scientists make better decisions, close the loop between data and the lab, and deliver real impact and success in reducing Cost, Time, and increasing the Quality of compounds which reach Phase 1, 2 and 3 Clinical trials.
The Platform Vision included a Levels of automation framework, outlining increasing levels of Automation from 0-5 - with Level 5 (Full autonomy) being the North Star. The as-is platform was primarily operating at the Level 2 (Connected workflows) level with some POCs launched at Level 3 (Partial autonomy) with the majority of platform products and tools falling under the Level 0 category (Tool level).
03 Challenge
The challenge was to help a cross-functional team consisting of Executives, Engineers, Medicinal Chemists, Computational Drug Designers, AI researchers make sense of a broad platform vision and understand what it meant for their specific component, their users, and their future direction.
The 2026-2029 platform vision created an ambitious direction for the wider ecosystem, but the team needed the space to interpret it together, build shared understanding and translate it into something practical, meaningful and actionable. Whilst individual teams had already begun exploring agent-based POCs across the DMTL ecosystem. However this activity was happening at a Team level, without a shared strategic frame for how these emerging agentic capabilities should connect back to the wider platform vision.
Another key challenge was to bring these parallel explorations together, creating shared visibility across the group, and translate distributed experimentation into a coherent set of functional capabilities and roadmap themes.
It was also crucial to ensure that different perspectives were heard, across Product, Science, Engineering and Leadership, whilst also guiding the group towards a clearer collective position. Without this alignment, there was a risk that sub-component teams would move forwards with different assumptions, creating fragmentation across the platform.
It was also crucial to ensure that different perspectives were heard, across Product, Science, Engineering and Leadership, whilst also guiding the group towards a clearer collective position. Without this alignment, there was a risk that sub-component teams would move forwards with different assumptions, creating fragmentation across the platform.
The work also needed to create confidence around where agentic systems could genuinely support future Scientific workflows. This meant moving beyond generic AI opportunity spotting and helping the team to identify where automation, decision support or adaptive workflows could reduce friction, improve scalability, and better support Scientists in their day to day work.
Ultimately, the challenge was to turn a broad strategic vision into a shared, practical direction that gave teams clarity, momentum and a stronger basis for future decision making.
My role was to bring people together around a shared direction. I created and facilitated the workshop structure that allowed different voices to be heard and move the group from broad ambition into clearer themes, opportunities and workstreams, particularly around where agentic systems could meaningfully support future Scientific workflows and increase operational efficiency. This included:
- Designing the workshop structure and facilitation approach
- Creating the conditions for open, constructive cross-functional collaboration
- Surfacing different stakeholder perspectives across Product, Science, Engineering and Leadership
- Guiding the group from discussion into clear priorities, themes and next steps
Agree and align on a shared future vision between Design, Predictive Chemistry, Protein and Atomistic Data Layer [PADL], Synthesis AI, Computational Drug Design, Quantum mechanics, Design Automation and DMTL [Design, Make, Test and Learn] Lab components in response to the Company Tech Platform vision.
Draft and review Platform goals shared between each component.
Agree and align on end of year statements across Design, Predictive Chemistry, PADL and DMTL
06 Approach
07 Key artefacts
Following the workshop, I consolidated the group inputs into a strategic synthesis deck that turned a large amount of discussion, team input and existing agent-based POC activity into a clearer future-state direction for the component. The synthesis deck translated the 2026–2029 platform vision into a more practical view of how the DMTL workflow could evolve across Design, Make, Test and Learn.The output included:
- Updated 2029 “to-be” DMTL future-state workflow
- Agent-supported capability themes across Design, Make, Test and Learn
- Human-in-the-loop checkpoints for expert review, approval and escalation
- Levels of automation shifts from current state to future state
- Value proposition and impact statements for each DMTL stage
- Risks, dependencies and assumptions to manage before delivery
- Roadmap inputs for future agent-supported product discovery
This gave the group a shared artefact to align around, helping teams move from parallel exploration into clearer product direction and downstream workstreams.
Artefact 4: Case study 1 - Levels of Automation
08 Strategic value
The strategic value of the work was to connect agentic platform ambition back to the outcomes that matter in scientific discovery: better programme quality, stronger novelty, greater scale and faster learning. Rather than measuring success simply by “more automation”, the platform vision needed to be evaluated by whether it could help teams progress better programmes, make higher-quality decisions and improve the discovery system over time.
Novelty: Prioritising programmes with stronger biological and patient insight, using internal, partner and open datasets to identify opportunities with a clearer reason to win.
Quality: De-risking programmes earlier by clarifying TPP profiles, stage-gate criteria and the most important risks to resolve at each stage.
Scale: Increasing the number of programmes moving through the platform, while improving the proportion that successfully traverse stage gates.
Continuous improvement: Optimising information gain across both individual programmes and the wider platform, creating feedback loops that improve future decisions and system performance.
09 What this demonstrates
This case study demonstrates my ability to:
- Translate an ambitious AI platform vision into practical product strategy
- Identify where agentic capabilities can support real workflows, not generic AI use cases
- Design and facilitate strategic workshops across complex stakeholder groups
- Balance automation ambition with human judgement, governance and scientific accountability
- Structure ambiguity across scientific, product and engineering domains
- Turn raw workshop input into clear synthesis, roadmap themes and product discovery direction
Made with Bullet