It’s wild to think that ChatGPT only launched at the end of 2022. In the space of just a couple of years, AI has evolved faster than anything else in digital — faster than mobile apps, faster than social, and arguably faster than the early web itself, which, being a digital veteran, I remember all too well (even before Google reshaped search).

And unlike most tech waves, this one hasn’t quietly bubbled away in the background. Everyone in pharma and life sciences — from medical writers and brand managers to clinical teams, regulatory reviewers and even senior leadership — has now had some sort of hands-on experience with generative AI.

The question is no longer “Should we be using AI?”
It’s now “How do we use it properly, safely, and strategically — without tripping over compliance or diluting scientific accuracy?”

This article looks at how AI has evolved since ChatGPT’s launch, how teams are using it across the sector, the impact of Google Gemini 3, and what we can expect as we approach the next major wave of change in 2026. Most importantly, it explores what all this means for agencies and life sciences organisations working together.

A quick rewind: ChatGPT’s early days

When ChatGPT first appeared, it felt like a clever toy — brilliant for sparking ideas, drafting notes, or reframing content. But it had major limitations: hallucinations, no real memory, limited reasoning, and no real understanding of scientific precision.

Fast-forward to today and we now have:

  • Multi-modal AI (text, images, video, PDFs, charts, code — all in one flow)
  • Models with live data access and stronger reasoning capabilities
  • AI woven into everyday tools like Teams, Office, Google Workspace, Jira and Notion
  • AI agents capable of running defined tasks, not just generating text
  • Enterprise-grade security, governance and auditability

In other words, AI has moved from being a novelty to becoming genuine infrastructure. For pharma, MedTech and wider life sciences, that shift is enormous.

What Google Gemini 3 means for life sciences

It’s not just OpenAI driving the pace. Google’s new Gemini 3 model introduces another step change — particularly relevant to regulated industries.

Gemini 3 is designed as a fully multimodal system that can analyse text, charts, slides, PDFs, structured data and images in a unified workflow. That’s incredibly useful in life sciences, where content routinely spans SmPCs, training decks, regulatory documentation, clinical summaries and large Excel sheets.

Because Gemini integrates directly with Google Workspace — Gmail, Docs, Slides, Sheets and Drive — AI will increasingly appear in the tools teams already use daily. Tasks like summarising regulatory PDFs, comparing PI versions, checking slide accuracy, and drafting training content become far quicker and more consistent across organisations.

Crucially, Gemini 3 also arrives with stronger enterprise controls: permission settings, access governance and audit trails — all essential for ABPI- and MDR-aligned environments.

We’re heading into a multi-model world, where both OpenAI and Google underpin different parts of the workflow depending on use case, risk profile and infrastructure.

How teams are using AI today

Across our pharma and life sciences projects, I’m seeing four major categories of usage — usually starting informally, then maturing rapidly once the time savings become obvious.

1. Content creation and medical writing (done properly)

Teams are using AI to support:

  • Writing and refining briefs
  • Drafting slide decks, value propositions, training materials and promotional content
  • Restructuring dense scientific text into clearer narratives
  • Highlighting inconsistencies across long or complex documents
  • Summarising SmPCs and literature (always with expert checks)

The key point is that AI is speeding up the thinking, not replacing expert judgement.

2. Coding and automation

This is one of the biggest shifts behind the scenes.

Developers now use AI to:

  • Generate boilerplate code and repetitive patterns
  • Prototype and test ideas much faster
  • Identify and debug issues
  • Automate workflows such as content logs, formatting and compliance prep

What used to take hours now takes minutes, freeing specialist agencies to focus on UX, logic, and regulatory alignment rather than repetitive build work.

3. Data analysis and insight generation

AI is transforming how brand teams interact with data. It’s now one of the fastest ways to interrogate huge Excel sheets, dashboards and RWE/CRM outputs.

Teams are using AI to:

  • Spot trends and behavioural patterns
  • Forecast and test scenarios
  • Compare performance across time periods or segments
  • Identify anomalies or outliers that need investigation
  • Turn numbers into clear, digestible insights for stakeholders

Paired with anonymised datasets, AI becomes a powerful, near real-time insight engine.

4. Day-to-day productivity

Then there’s the unglamorous but genuinely transformational work:

  • Drafting emails and internal updates
  • Preparing agendas and meeting packs
  • Turning notes into structured tasks and follow-ups
  • Rewriting complex regulatory text into clearer language
  • Planning content calendars and campaign timelines
  • Supporting launch preparation and internal training

AI has quietly become the colleague everyone relies on — and yes, many of us now say “please” and “thanks” to it.

How this is reshaping agencies (and why clients are noticing)

The traditional agency model — large teams, long turnarounds, layered account management — simply doesn’t fit an AI-powered world. Specialist life sciences agencies become even more valuable because AI amplifies expertise, not headcount.

Time is being redeployed, not removed

AI is stripping out the heavy lifting behind:

  • Research and information gathering
  • First drafts and initial concepts
  • Formatting and content structuring
  • Compliance preparation and cross-checks
  • Code scaffolding and technical boilerplate
  • Document conversions and reformatting

That frees agencies to focus on:

  • Strategy and positioning
  • UX and design for HCPs and patients
  • Scientific clarity and accuracy
  • Compliance alignment and ABPI/MDR sensitivity
  • Problem-solving across digital ecosystems
  • Innovation in workflows and platforms

It’s a shift from “deliverables” to genuine thinking partnerships.

Clients expect speed — but not at the expense of accuracy

AI has raised expectations, and rightly so. People know things can be done quicker.

But in our world, faster must still mean:

  • Scientifically accurate
  • ABPI- and MDR-aligned
  • Technically robust
  • Compliantly executed and documented

Agencies that deeply understand med-legal workflows and regulatory-grade publishing are the ones that will thrive.

More collaborative working

AI is also flattening the agency–client dynamic. We’re seeing:

  • Shared prompts and prompt libraries
  • Real-time co-drafting sessions
  • Faster decision-making cycles
  • More collaborative planning workshops

The result? Less friction, better outcomes, and stronger long-term partnerships.

What to expect in 2026

We’re entering a phase where AI becomes the operational engine behind digital work — not just a helpful assistant on the side.

1. AI agents performing real operational tasks

By 2026, AI agents will be capable of handling defined, governed tasks such as:

  • Updating digital PI and checking for consistency
  • Comparing PDFs against SmPCs and flagging deviations
  • Running accessibility checks across websites and apps
  • Flagging inconsistencies in medical and promotional content
  • Preparing first-pass ABPI-aligned materials for human review
  • Monitoring AE-related content pathways and touchpoints

In other words, tireless, compliant junior teammates that never get bored and always document what they do.

2. Hyper-personalised digital experiences

We’re moving beyond traditional “dynamic content” towards experiences such as:

  • Adaptive HCP dashboards that respond to speciality and behaviour
  • Dynamic PI via QR codes, always pointing to the latest information
  • RWE feeds tailored to indication, region or patient population
  • Micro-sites that adapt to how users interact, not just who they are

All of this can be delivered without breaching privacy lines, provided the data strategy is designed correctly.

3. AI-ready websites and platforms

The next generation of sites and platforms will be built with AI in mind from day one:

  • Structured, AI-readable content instead of free-form blobs
  • Modular, component-based blocks that can be reused intelligently
  • Embedded data models to give context to content
  • Automated compliance alerts when content drifts off-label or out-of-date
  • Smarter HCP/patient split logic to protect the right audiences from seeing the wrong content

This all massively reduces compliance risk and the burden of manual updates.

4. Agencies and life sciences teams operating as blended units

The relationship is shifting from:

“Give us the brief and we’ll deliver.”

to:

“Let’s design the workflow together — and let AI take care of the repetitive bits.”

Agencies with regulatory, technical and automation expertise become operational partners rather than just production partners.

Where MCP fits in — and why it will be the buzzword of 2026

If there’s one technical concept set to dominate 2026 conversations, it’s the Model Context Protocol (MCP).

MCP is an open standard that lets AI connect to external systems — CMSs, CRMs, analytics platforms, regulatory libraries — in a secure, permissioned and fully auditable way.

It’s the missing connective tissue between AI and the systems life sciences teams depend on.

Why MCP matters:

  • AI can only access what it’s explicitly allowed to
  • Every action is logged and auditable
  • No uncontrolled data flows or shadow integrations
  • No risky, one-off API hacks
  • Perfect for ABPI- and MDR-heavy environments

MCP elevates AI from “assistant” to safe operator.

For agencies like ours, it enables AI-ready ecosystems where PI updates, approvals, compliance checks and content changes can run automatically — without tearing up or rebuilding the underlying infrastructure.

MCP won’t be “just another acronym”. It is likely to redefine how pharma, MedTech and wider life sciences manage digital content over the next five years.

If the last two years were about experimenting with AI, and 2025 is about embedding it, then 2026 will be about connecting it — safely, securely and strategically.