Prove they can do it in Excel.
Set an Excel exercise. Trainees download the starter workbook, complete it in Excel, and upload it back. MLtitude grades against the rubric you set — cells, formulas, sheet shape, formatting — and writes the trainee personalised feedback. No backlog. No grading.
“They completed the training” isn't a measure.
Excel is the second job description in finance, operations, analytics, sales-ops, FP&A — anywhere a spreadsheet is real work. L&D teams run training on it constantly. Most of the “did they learn it” measurement is still self-report and a multiple-choice quiz.
- LMS marks the module complete. We assume they watched it.
- Multiple choice at the end. Four options. They pass.
- Manager asks the team six weeks later — “Are you using it?”
- Three trainees actually changed their workflow. The others didn't.
- Next quarter we run the same training.
- Trainee downloads the workbook. Real Excel, real data.
- They complete the task. We grade against the rubric.
- Per-check feedback tells them what to fix and why.
- Trainer sees the cohort: who's stuck on what, who's done.
- Mastery — not seat-time — becomes the measurement.
The programmes that need proof.
Anywhere a spreadsheet does real work, “they completed the training” stops being enough. These are the four programmes XLSim earns its seat on.
From the model to the muscle memory.
Internal models, recons, reporting workflows. The new joiner doesn't just read about VLOOKUP — they prove they can do it on data that looks like the data they'll see on Monday.
Pivots, lookups, and SUMIFS — verified.
The skills you'd want before someone touches the live dashboard. Demonstrated against a workbook, not promised in a self-rating.
CRM export → pipeline analysis.
Show the work, not just the click-through. The output is a workbook that does what it claims to do, graded automatically and returned with feedback.
The test they passed, not the deck they watched.
Documented competency for the audit, not 'they were marked complete'. XLSim records the attempt, the score, and the per-check trail — exportable when finance auditors come asking.
How an assessment actually runs.
Two roles. One workflow. The trainer authors and assigns; the trainee attempts and receives feedback. The engine handles the grading and writes the narrative. No spreadsheet inspection, no marking up a workbook by hand.
Pick from the seeded library (Fundamentals, Intermediate, Advanced) or generate a new assignment from a brief. Edit the rubric. Set the passing score.
Pick the trainees — one, a cohort, the whole programme. Idempotent, so you can re-assign without duplicating.
They see only what's assigned. Download the starter, complete in Excel, upload the .xlsx. No XLSim-specific app — they use the tool they already know.
The engine runs the rubric against the workbook — cells, formulas, ranges, sheets, formatting. Each check returns pass/fail with detail.
A short narrative explains what passed and what to revise. Trainer sees the cohort view: scores, attempts, who's stuck where.
Cell value
A specific cell must equal a specific value.
Cell formula
A cell must use a specific formula shape.
Required functions
The formula must contain certain function names.
Range values
An entire range must contain the right sequence.
Sheet exists
A specific sheet (by name) must be present.
Formatting
Cells must use a specific format — currency, bold, fill, font size.
Composable. Build a rubric of any size by mixing check types. Grading is deterministic — a model writes the per-check narrative, but the score itself is calculated, not opinion.
The questions L&D leads ask.
Procurement, security, and brand teams ask their own questions too. If we missed yours, talk to us — we read every message.
Is the grading actually accurate?
Yes — and it's deterministic. The engine runs each rubric check against the workbook (e.g. “D6 must equal 74000”, “C2's formula must contain IF”). No model is making the pass/fail call. Where a model is involved is the per-check narrative; the trainer can override any score.
Do trainees need to install anything?
No. They open the assigned assignment in their browser, download the starter .xlsx, complete it in whatever Excel they already have (desktop, online, LibreOffice — anything that produces a valid .xlsx), and upload it back. We meet them in the tool they already know.
How do we author the rubric — is it technical?
Two ways. (i) Pick from the seeded library — Excel Fundamentals, Intermediate, Advanced — and assign as-is. (ii) Generate a new assignment from a brief: describe the exercise in plain English and the system writes the title, instructions, rubric, and starter workbook. Either way, you can edit the rubric before publishing.
Can trainees re-attempt?
Yes, by default. You can also cap attempts per assignment. Every attempt is stored separately with its score, evaluated_at timestamp, and per-check trail — useful for the audit conversation later.
Will it work for our internal Excel templates?
Yes. The starter workbook can be your internal template (we host it; trainees download it). The rubric checks the parts that matter for the exercise — typically not the whole template, just the cells the trainee was supposed to fill.
What gets logged for audit?
Every submission — file, attempt number, score, max score, pass/fail, evaluated_at, per-check results, and the rubric used. Org admins can export the workspace audit log as CSV.
Run one assessment. Watch a cohort.
Pick a real Excel skill you're training this quarter. Assign it to one cohort with XLSim. Watch what happens when “did they learn it” stops being a self-report. Then decide whether you'd run every Excel programme this way.