From Scattered AI Use to Five Working Systems

The Challenge

BIMA operates across clinical support, advocacy, communications, leadership and events — powered largely by volunteers. AI use was inconsistent and unstructured:

• Knowledge retained by individuals, not the institution
• Complex cases manually rewritten and summarised by hand
• Inconsistent communications across channels
• Varying planning processes between teams
• High cognitive burden on the staff and volunteers drafting content

The Approach

The engagement combined foundational AI education, one-to-one implementation sessions, custom AI assistant design and build, institutional knowledge structuring, and a consolidation showcase.

The Systems Built

1. Support Clinic Assistant — Clinical Operations

Processes complex case material from messy inputs, generates clinician-style reports and case summaries, anonymises outputs and flags missing information. Hours of summarising reduced to minutes.

2. Event & Project Planning Assistant — Internal Operations

Converts notes into structured project plans, extracting risks, timelines and decisions. Handovers between teams became dramatically clearer.

3. Communications & Marketing Assistant — External Communications

Generates audience-specific content in brand tone: newsletters, social posts and member communications — producing 80–90% usable drafts from a single brief.

4. Executive Personal Writing Assistant — Leadership

Drafts speeches and LinkedIn posts, and converts internal updates into public communication. "Already 70 percent there" after first testing.

5. Advocacy Knowledge Assistant — Policy

Stores institutional memory and historical positions, drafting structured policy responses with terminology guardrails. Knowledge shifted from individual memory to institutional architecture.

The Outcomes

• Multiple functioning custom AI assistant systems live and in use
• Structured internal knowledge bases established
• Clear departmental ownership of each system
• AI embedded in live workflows — not a side experiment
• Increased confidence in responsible AI use

In Their Words

"Previously taking one to two hours for case note condensing — now taking minutes. Very straightforward. Brilliant." — Clinical Operations team member

"More structured than what we normally have — it could be sent straight to the team." — Internal Operations lead

The Communications team applied their system to website content and emails within one week of implementation.