Read More
10 Min Read
AI Won’t Make Bad MSL Insights Better: But Here’s What Will
AI is blowing up the MSL insights world, and rightfully so. One of the biggest pain points in the MSL insights process is analyzing large volumes of free-text insights. It can be overwhelming, slow, and one of the biggest bottlenecks in making insights actionable.
GenAI changes the game. It can take a huge pile of insights, summarize them, and pull out patterns and trends in a fraction of the time. It’s fast, efficient, and scalable.
But there’s a major misconception: that AI can take bad MSL insights and magically make them better.
I’ve been hearing VPs of Medical Affairs say, “Oh, AI will take care of the bad insights.”
This is not the case. AI cannot take incomplete or vague insights and transform them into decision-grade insights that drive strategy.
To get you thinking about this more deeply, I ran an experiment. It highlights something every Medical Affairs leader and MSL needs to know:
👉 AI doesn’t replace the need for strong insights training. It makes it even more essential.
Here’s why AI won’t make bad MSL insights better:
How Large Language Models (LLMs) Work (at a High Level)—And Why They Can’t Fix Bad Insights
Before we dive into the experiment, let’s take a moment to understand how large language models (LLMs) actually work.
At their core, LLMs are sophisticated pattern-recognition and summarization tools. They analyze vast amounts of text, predict the most likely next word in a sequence, and generate responses based on probabilities learned from massive datasets. This makes them excellent for:
✅ Summarizing large volumes of text
✅ Identifying common patterns and themes
But here’s the catch: LLMs don’t generate new meaning where none exists.
This is critical for MSL insights. If an insight is incomplete or missing context, AI can’t magically fill in the missing gaps. If the original insights lack depth, AI will simply amplify that lack of depth.
So while AI can accelerate analysis, it cannot transform incomplete insights into strong, decision-grade insights. That’s why MSL insights training is more important than ever.
The AI Experiment: Can It Fix Bad MSL Insights?
To test this assumption, I ran a simple experiment using AI. I used ChatGPT-4o for everything described.
Step 1: Generate Bad MSL Insights
First, I had AI generate 100 bad MSL insights. The kind of insights Medical Affairs leaders complain about:
🚫 “The HCP liked the Phase 3 data.”
It’s a the start of potentially exciting insight. But it’s lacking:
❌ Context: Where did this insight come from and who said it?
❌ The Why: What’s driving the HCP’s opinion?
❌ Implications: How does this impact strategy, the business, or patient care?
Step 2: Rewrite the Insights to Include Context, the Why, and Implications
Next, I asked AI to rewrite those insights to include the missing pieces, context, the why, and implications, making them more comprehensive.
Example of a rewritten, more comprehensive insight:
✅ “The HCP appreciated the Phase 3 data because it showed a 25% reduction in hospitalization rates compared to the standard of care. They believe this could address a significant unmet need for high-risk patients and improve adherence rates due to fewer hospital visits.”
Step 3: Have AI Analyze Both Sets of Insights
Here is the prompt I used to analyze both sets of insights (independently):
“You are an expert in analyzing msl insights and identifying trends and decision grade insights. your task is to analyze 100 insights and create a report of the major trends. here are the insights: [Insert insights]”
The Results: With Complete Insights, AI Suggested Better Recommendations
When AI analyzed the incomplete insights, it made some suggestions, but they definitely wouldn’t wow anyone in Medical Affairs. The recommendations were generic, surface-level, and lacked strategic depth.
Interestingly, AI did recognize the need for deeper insights and suggested gathering more detailed, context-rich insights (probably because it’s reading my context history). Here’s are the recommendations:
These are reasonable recommendations, but they’re nothing you would be excited to show to your boss. The AI generated insights lacked depth, and true strategic value, which makes sense, because the data itself was vague and shallow.
What Happened When AI Analyzed the Complete Insights?
In contrast, when AI analyzed the more complete insights, the recommendations became significantly more detailed, and actionable. The analysis suggested:
The Takeaway: AI Won’t Fix Bad MSL Insights
Obviously, this experiment isn’t perfect and has plenty of limitations.
But the point remains crystal clear: AI won’t fix bad insights.
🚨 Garbage in = garbage out
🚨 If the input lacks depth, the recommendations will lack depth
🚨 If the insights aren’t comprehensive, the AI won’t magically uncover decision-grade insights
AI can be a game-changer for insights analysis, but only when it’s working with high-quality, complete insights. That’s why MSL insights training is more important than ever.
What are your impressions of these two analyses? Do you want to play around with analyzing these insights in AI? Maybe you are a vendor with a special AI model for MSL insights. Ping me, I am happy to share and compare analyses (in fact, I would love to!).
Reach out to me to share your thoughts!
Why MSL Insights Training Is More Important Than Ever
AI is an amplifier, not a fixer. If your team is collecting vague, incomplete insights, AI will only highlight the lack of strategic value in those insights. It won’t fix them.
To fully leverage AI, MSLs must be trained to capture high-quality, comprehensive insights.
1. Training MSLs to Gather Complete Insights
High-quality MSL insights are built on three pillars:
🔹 Context: What is the situation in which this insight was gathered?
🔹 The Why: Why does the HCP feel this way?
🔹 Implications: How does this insight impact strategy, patient care, or business outcomes?
Learn more in my MSL Insights Online Course with a monster module on how to gather comprehensive insights.
Or reach out about my team trainings on insights. We can design a fun training that works well with your team culture.
How to Improve MSL Insights Gathering
MSLs can gather better insights with training on asking better questions and how to use key questions for insights. Examples include:
❓ “What aspect of the Phase 3 data do you find most relevant to your practice?”
❓ “Why do you believe this safety profile will encourage broader adoption?”
❓ “Tell me more.”
Here are some additional resources on how to ask better questions. Also reach out about my team training on questioning techniques.
GET THE CHEAT SHEET
MSLs need to have many ways to uncover KOLs motivations without acutally using the word, “why.” Here’s how.
MSL OMNICHANNEL INSIGHTS
Understand your HCPs’/KOLs’ communication preferences. Use this to have better interactions and gather great insights.
23 QUESTIONS FOR MSLS
Download this 1-pager with 23 questions to keep in your back pocket and ask great questions that uncover the why.
2. Training MSLs on How to Use AI for Pre-Meeting Preparation
AI is a powerful tool for pre-meeting planning, helping MSLs:
✅ Identify key HCP/KOL interests
✅ Uncover trends in therapeutic areas or emerging data
✅ Anticipate questions HCPs may ask and generate useful questions to uncover insights
By entering meetings with a clear plan, MSLs can focus conversations on gathering insights that matter.
AI isn’t going away. Don’t let your team fall behind. Learn more my AI for MSL excellence training to help your team use AI in their day-to-day work and get comfortable with it.
Conclusions: AI Won’t Make Bad Insights Better
This experiment proves what Medical Affairs leaders need to understand:
📌 AI doesn’t replace the need for strong MSL insights training—it makes it even more essential.
📌 If your MSL insights are incomplete, AI will only make that more obvious.
📌 To unlock AI’s full potential, Medical Affairs teams must train MSLs on how to gather complete, high-quality insights.
For Medical Affairs teams to truly succeed in the AI era, they must invest in:
✅ Training MSLs on gathering complete insights
✅ Training MSLs on how to use AI effectively
When MSLs master the art of capturing context, the why, and implications, AI can amplify their impact turning insights into actionable strategies that improve patient outcomes and demonstrate the value of Medical Affairs.
Read Next

Why Course Correction is a Must for Stakeholder Engagement Strategy
5 Min ReadWhy Course Correction is a Must for Stakeholder Engagement StrategyViews expressed in this article are personal. Imagine you're the captain of a ship. The sea is vast, the weather unpredictable, and the destination is clear but far away. You’ve got...

MSL Omnichannel Insights: Why They Are Important and What to Focus on
Omnichannel is probably the most talked about topic right now in Medical Affairs. This was a huge focus area at the recent Medical Affairs Professional Society 2023 meeting in Nashville. There is so much buzz around omnichannel and rightfully so! Omnichannel is an...

6 MSL Insights Truths I Wish I Knew 7 Years Ago
10 Min Read6 MSL Insights Truths I Wish I Knew 7 Years AgoSeven years ago, I stumbled into the world of MSL insights by complete chance. I didn’t have a roadmap or even a clear understanding of what insights really meant for Medical Affairs. I went from knowing...
0 Comments