Monday, nine o'clock. The launch dashboard is green from top to bottom. Every rep in every country has opened the modules, watched the product video and passed the end-of-course quiz. Completion: 100%. Then the first ride-along reports arrive, and they tell a different story. One rep can't explain how the new device compares with the incumbent. Another quotes the old indication wording. A third handles the price objection by changing the subject.
Nothing on the dashboard was false. It measured the wrong thing. Completion tells you that training was delivered. It says almost nothing about what each person now knows, and even less about what they will still know in week six. This piece is about closing that gap: what we mean by knowledge intelligence, what data it adds to what many training systems already hold, and how to build it into a launch or a cycle meeting without adding training days to anyone's calendar.
Completion is a delivery metric wearing a readiness costume
Most pharma and MedTech organisations already run a learning management system, and they should. It hosts the onboarding curriculum, the compliance modules, the approved product training and the certification records that auditors ask for. It does that job well.
The problem starts when the completion report is asked to answer a question it was never built for: is the field ready? A completed module means someone reached the last screen. An end-of-course quiz, taken minutes after the content, mostly measures short-term recall. Both happen on the day of delivery, which is exactly the moment when knowledge looks best and is least representative of what will survive.
Murre and Dros's 2015 replication of Ebbinghaus' forgetting curve, a single-participant experiment with nonsense syllables, illustrates forgetting in a specific experimental task. It can't tell you how fast product knowledge fades, but it is reason enough not to read a quiz passed on Tuesday as a forecast of the conversation in a respiratory clinic three weeks later.
There is a second, quieter problem. Completion tells you who met the module's requirements. It says nothing about what they kept. And it is almost always high in regulated industries because training is mandatory. So the metric that fills the dashboard has almost no variance. It can't tell the strong territory from the weak one, the rep who understands the mechanism from the rep who memorised a slide, or the country where the label change landed from the country where it didn't.
Knowledge intelligence means knowing who is wrong about what, and whether it gets fixed
We use the term knowledge intelligence for something quite specific. It is the ability to see, person by person and question by question, what your people actually know about the things that matter commercially, to act on each gap individually, and then to verify whether that gap closed.
Three words in that definition carry the weight.
- Person by person, because a team average of 82% can hide a territory where half the reps have the dosing wrong. Readiness is only useful when it can be broken down by individual, team, country and topic.
- Act individually, because if one rep misunderstands the primary endpoint and another confuses two competitor devices, sending both the same refresher module wastes time for both.
- Verify, because reading the correction proves nothing until the same person gets it right later.
In practice, we define a knowledge gap as one person answering one question wrong at their first attempt. It counts as closed only when that same person answers it correctly on a later attempt. That definition is deliberately strict. It stops you from celebrating the fact that the team average went up because new, stronger players joined, and it forces a conversation about the gaps that stayed open.
A gap is not closed because someone read the correction. It is closed when the same person, asked again later, gets it right.
Where typical LMS set-ups fall short, and what the layer adds
A fair objection is that most LMS platforms already have quizzes, question banks and reports, and many teams already run tests and reinforcement. Why not just use them more?
Often you can, and some teams do it well. But in the set-ups we see most, three features of how the LMS is used get in the way.
First, it is built around courses. People go in, consume, get assessed and leave. Measurement happens at the end of an event, not continuously in the weeks after it, which is when decay and drift occur.
Second, sessions are long. A field rep between two calls, or a key account manager waiting outside a hospital pharmacy committee, has four minutes, not forty. Anything that needs a laptop, a login and a quiet room will be done on Friday evening, if at all.
Third, few people go back unprompted. For many reps the LMS is where you go because you have to. That is fine for compliance. It is hard on reinforcement, which works through repeated retrieval spread over time.
So we see knowledge intelligence as a layer next to the LMS, defined by two things it adds. The data: for every person and every question, the first-attempt answer, which wrong option they chose, every later attempt, and whether the question is flagged critical. And the follow-up workflow: each wrong answer triggers that person's own fix, the question returns on a schedule, closure is judged against a fixed threshold, and whatever is still open after a set period goes on a weekly named list to the manager, or to Medical if it touches the label.
What the layer looks like in practice
This is how we build it in Atrivity, and the principles apply whatever tool you use.
Short retrieval, not more content
The core unit is a one-minute challenge: five questions in sixty seconds, played against real colleagues. It fits in the gap between two visits. It forces retrieval rather than rereading, and retrieval is what strengthens memory. Content still matters, and each question links back to a short, approved piece of material, but the engine is the question, not the slide.
Competition that brings people back
Reps are competitive. Scoreboards, team battles and a synchronised window around a kickoff or a cycle meeting generate a kind of voluntary repetition that mandatory modules rarely do. The point of the game is not entertainment. It is volume: enough answers from each person, spread across enough days, to see which gaps close on their own and which do not.
Individual fixes, sent automatically
When someone gets a critical question wrong, they receive the fix for their own mistake, not a generic recap. We call it Gap Fixing. The question comes back later, and a critical gap counts as closed only after 5 of 5 later check questions on the same concept answered correctly. One lucky guess is not proof.
A readiness view that managers can act on
The output is a readiness index by team, country and topic, and a list of the specific misconceptions that are still open. That second part matters more than the index. A manager going on a ride-along next week needs to know that four of her seven reps still confuse the two patient populations in the trial, not that her team scored 78%.
Where this pays off first
We see the value of knowledge intelligence most clearly in five situations. They have one thing in common: a short window in which the field has to absorb something new and use it in front of customers.
- Launches. The first year sets the trajectory. McKinsey's analysis of 210 launches between 2003 and 2009, published in 2014, found that about two thirds miss pre-launch consensus expectations in year one, and that 78% of year-one underperformers were still underperforming in year two. Every week in which the field tells the story inconsistently is a week you are unlikely to win back.
- Kickoffs and cycle meetings. The POA sets new messages, new segments and sometimes a new visual aid. The meeting creates energy. The weeks after it decide whether the messages survive contact with the territory.
- Label and indication changes. When the approved wording changes, the old wording lingers for months unless you can see who is still using it.
- Multi-country consistency. The same content in ten languages does not guarantee the same understanding in ten markets. Readiness by country shows where the local adaptation lost something.
- After the workshop. A two-day training event is expensive, and without reinforcement you have no way of knowing how much of it survived.
The gaps that won't close on their own
Here is the finding that surprised us most in the programs we run, and it is the strongest argument for measuring at the level of the individual question.
Some gaps receive a verified correction after repetition; others remain open and need further assessment or support. Someone gets a question wrong, sees the correct answer and the short explanation, meets the question again a few days later and gets it right. That is the forgetting curve being pushed back, and a well-designed reinforcement layer does it at scale without anyone having to organise it.
But some gaps barely move. They are not forgotten facts. They are misconceptions: a wrong mental model of the mechanism of action, a confusion between two trial populations, a belief about a competitor that the rep has repeated for years. Repetition does little against them because the person is not missing information. They are holding the wrong information with confidence.
Those gaps need a different response. A conversation with a manager, a discussion with an MSL, a coaching moment on a ride-along. The job of knowledge intelligence is to find them early and name them precisely, so that scarce coaching time goes where repetition has already failed. Without that separation, managers coach everything a little, and the stubborn misconceptions survive into the second quarter of the launch.
Picture week three of a launch. The readiness view shows that most gaps on the new efficacy data are closing nicely. One question, about which patient population the data applies to, is stuck: a group of reps keeps choosing the broader population, attempt after attempt. That is not a knowledge problem you solve with another quiz. It is a compliance risk and a credibility risk, and it goes straight to the field managers and the medical team with names attached.
What changes for the people involved
For Commercial Excellence and Launch Excellence, readiness stops being an opinion formed in the launch meeting and becomes a measure you can track weekly, alongside call activity and early prescribing signals.
For first-line managers, coaching gets specific. Instead of "brush up on the product", the conversation is about the one thing a particular rep keeps getting wrong.
For Medical Affairs, the list of open misconceptions is an early warning. If the field misreads a secondary endpoint, it is better to find out from the data than from an HCP complaint.
For L&D, the question changes from "how much training did we deliver?" to "how much of what matters did we fix?", with the caveat that a fix verified in the assessment is not yet proof of what happens in the call.
And for the reps themselves, a five-question challenge between visits is often less intrusive than another hour-long module.
Start with one moment and one list of questions
You don't need a new learning strategy to begin. You need one business moment and one honest list.
- Pick the next moment that matters: a launch wave, a kickoff, a cycle meeting, a label update.
- With Marketing and Medical, write down the twenty or thirty facts the field must get right in front of customers. Not everything in the training deck, only what changes the conversation.
- Turn them into questions with plausible wrong answers, drawn from the approved material, and put them through the usual medical-legal review.
- Measure at first attempt, reinforce for several weeks, and verify closure person by person.
- Take the list of gaps that did not close to the managers and the medical team, by name and by topic.
That is what we set up with clients in Atrivity: the challenges, the individual fixes, the readiness view and the list of stubborn misconceptions, sitting next to the LMS rather than instead of it. If you want to see how it would look for your next launch or cycle meeting, we are happy to walk you through a demo built around your own situation.
And if you only take one question into your next launch meeting, make it this one: which three things is our field most likely to get wrong, and how would we know?