Proficiency Features require granting Pragmatico API Access to your AI tenant.
Analytics – Proficiency Scorecard
The Scorecard is your organization's monthly AI proficiency report: one score out of 100, the two dimensions behind it, how your people are spread across five tiers, and which kinds of work they are actually using AI for. You get there from Scorecard, under AI Proficiency in the left-hand navigation.
Only admins can open it. Everything on it is organization-level, not your own usage.
The score and the tier that goes with it
Org AI Proficiency Score is one number out of 100, described on the card as "One score combining Quality and Usage into your organization's overall AI proficiency". The five tiers are fixed to the score:
Score | Tier |
80 and above | Power User |
60 to 79 | Proficient |
35 to 59 | Developing |
15 to 34 | Dabbler |
Below 15 | Lagging |
Under the number is a short read of what the tier means, and a second line describing your own distribution. The bar at the bottom of the card marks where your score falls on a 0 to 100 track.
Select Learn more to open the full breakdown: the score, its change against the last fully scored month, the tier spread, and the Quality and Usage detail in one view.
Quality and Usage
The Proficiency card carries both dimensions, each with a dial and three sub-measures.
Quality asks "Are participants picking AI up and using it well?"
Breadth: is the variety of ways someone uses AI, like chat, file uploads, connected apps, and different kinds of tasks. It rewards range, not frequency.
Depth: how far someone takes a single conversation, refining and building on it instead of stopping at the first answer. It rewards sustained engagement rather than more usage.
Sophistication: how thoughtfully someone works with AI- giving it context, checking its answers, and pushing back when something seems off. It rewards the craft, not the usage.
Usage asks "How often and broadly is AI used day to day?"
Frequency: how often someone turns to AI day to day and week to week. It rewards regular use and is not moved by occasional bursts of activity.
Volume: how much someone uses AI overall, in messages or conversations over time. It rewards sustained use and is not moved by sporadic big spikes.
Consistency: how evenly someone's use is spread over time rather than clustered into just a few days. It rewards a steady, repeating habit.
The two dimensions are measured on different scales. Quality is graded against a fixed standard, so a Quality of 70 means the same thing in June as it does in September. Usage is a position against everyone scored that month, so about half of any population sits below 50. A Usage score can fall because other organizations used AI more, not because yours used it less. Check your activity counts before reading a Usage drop as a drop in behavior.
What Needs Attention, Fair, Good and Optimal mean
Every score on the page, dimensions and sub-measures alike, carries one of four ratings:
Score | Rating |
90 to 100 | Optimal |
70 to 89 | Good |
50 to 69 | Fair |
Below 50 | Needs Attention |
Performance Tiers
This card answers "How is the team distributed across performance tiers?" It shows the percentage and headcount in each of the five tiers, with the weakest on the left. Each tier name carries an information icon explaining what that group needs, which is the most useful thing on the card: Lagging needs direct intervention, Dabbler needs re-engagement, Developing needs reinforcement and guided use cases, Proficient is ready for stretch work, and Power Users are ready to help scale AI across the organization.
When anyone is in Dabbler or Lagging, the pop-out names how many and offers a link straight to the Leaderboard filtered to exactly those people. That is the fastest route from a number to a list of who needs help.
User Behaviors
"Which AI use cases are users actually doing?" This is the share of your scored people applying AI to each kind of work, ranked highest first, across categories like Planning & Scoping, Data Processing, Research & Analysis, Meeting Prep & Follow-up, Content Generation, Code & Technical, and Summarization, with Other for anything uncategorized.
AI Proficiency Trend
This card answers "How is quality and usage progressing?" It plots your Quality and Usage scores month by month. Look at the two lines separately. When Quality rises, your people are using AI more carefully. When Usage rises, they have moved further ahead of everyone else scored that month. The two answer different questions, so one can climb while the other dips without either being wrong.
Filtering the page down
The month selector picks which scored month the whole page describes. Only fully scored months appear.
The department filter narrows every card to the people in the departments you pick. It appears only if your organization has departments set up.
Filtering by department hides the written analysis. The narrative paragraph under the score and the Learn more pop-out are whole-company by design, so both switch off when you apply a department filter, and Learn more reads as coming soon. Clear the filter to get them back.
Who is counted
The Scorecard averages the people who were actually scored that month, not your whole seat count, which is why the headcount here is usually smaller than your active user count.
Someone is left out rather than scored zero when there aren't enough conversations to grade. A person needs at least five gradeable conversations in the month, and their activity has to be on a connected enterprise AI account.