How does a medical communications agency speed up scientific work while protecting sensitive data? This customer story shows how Cactus Life Sciences uses Microsoft 365 Copilot and more than 30 custom agents to automate labor-intensive steps, while leaving human review intact. The company reports that structured data extraction is about 35% to 50% faster. Read the story to learn from Cactus Life Sciences' experience.
How is Cactus Life Sciences using Microsoft 365 Copilot in medical communications?
Cactus Life Sciences (CLS) uses Microsoft 365 Copilot to
reimagine document-heavy scientific workflows across scientific writing and project management.
Here’s how it fits into day-to-day work:
- Custom automation agents for key workflows
CLS built more than 30 custom agents using Microsoft’s agent builder framework. These agents are aligned to specific steps in clinical, regulatory, and scientific content workflows.
- Structured data extraction and literature review
Agents automate the retrieval and structuring of information from scientific literature, so writers can spend more time on deeper analysis and synthesis instead of manual data handling.
- Quality and compliance support
Additional tools help with abbreviation checks, formatting consistency, and alignment with regulatory and publication standards, while still keeping human review in place.
- Project management and collaboration
Project managers use agents to summarize email threads and generate task lists, which helps them manage large volumes of communication across distributed teams.
CLS’s approach is to let AI
support the labor‑intensive parts of the process, not own the task. Workflows have been decomposed and rebuilt so that Copilot handles repetitive steps, while people focus on higher‑value scientific judgment and client outcomes.
What measurable impact has Copilot had on efficiency and collaboration?
CLS reports clear, measurable gains from its Copilot rollout, especially in structured data work and day-to-day collaboration.
Key outcomes include:
- Faster structured data extraction
Structured data extraction is now completed approximately 35–50% faster, while outputs still go through established human review and governance.
- More time for higher-value work
By automating search, extraction, and formatting tasks, scientific writers can focus more on interpretation, synthesis, and scientific quality instead of manual processing.
- Smoother project management
Project managers use agents to summarize email communications and auto-generate task lists, which helps them manage overflowing inboxes and coordinate remote teams more effectively.
- Peer-led innovation culture
A Copilot Champions program, launched about 6–8 months before this case was documented, has created a community that shares prompts, agent ideas, and best practices. This peer-led model is driving experimentation and new use cases across the organization.
Overall, Copilot is helping CLS
reshape how work gets done—not just by automating tasks, but by encouraging teams to rethink workflows from the ground up while keeping quality and governance intact.
How does Cactus Life Sciences keep AI use secure and human‑anchored?
Security and human oversight are central to how CLS uses Microsoft 365 Copilot.
Here’s how they manage it:
- Enterprise-grade security as a baseline
Security was a non‑negotiable requirement. CLS chose Microsoft 365 and Copilot because of enterprise-grade authentication, built-in protections, and the ability to work in dedicated project environments for sensitive client data.
- Human Anchored AI approach
CLS maintains a Human Anchored AI model: AI supports the work, but people remain involved at every step. All AI-enabled outputs are subject to the company’s established human review, quality control, and governance processes.
- Controls against hallucinations and bias
Quality control and governance processes have not been relaxed. They remain in place to address concerns about AI hallucinations, bias, and error tolerance, ensuring that scientific and regulatory standards are met.
- Structured adoption and training
A dedicated training team delivers modular programs so employees understand both the capabilities and limitations of AI tools. Adoption is largely peer-led through the Copilot Champions community, building on existing trust rather than relying solely on top‑down mandates.
CLS’s ongoing focus is to use AI to
support people, not replace them, and to keep evolving its use of Copilot in a way that protects client data while delivering sustained value in medical communications.
.jpg "FANTINI FABRIZIO"){kind=logo}
Sign in with LinkedIn.jpg)