How Crediteris works

The catalogue has four levels: Foundation, Practitioner, Advanced and Expert. Each level ends in a capstone case that certifies it. Pathways are mapped to four indicative role levels, R1 to R4. Your institution maps its own job grid to them during implementation.

Every role level has two pathways. Onboarding is for people joining from outside, who may know credit but not your mandate, your memo format or Canadian security law. Upskilling is for people already in the role who are moving up a level. An eighth pathway keeps certified senior bankers current each year.

A pathway works like this:

1. Every learner starts with the placement diagnostic, which estimates proficiency in each of the eight domains.
2. Fixed modules are required for everyone on the pathway.
3. Prescribed modules are added only where the diagnostic shows a domain below the target for the level. A learner who is strong in a domain skips them.
4. A learner can test out of any module except a capstone by scoring 85% or more with no critical error.
5. Capstones are never skipped. They are the certification evidence.

These rules run without AI. Pathways can also be opened for a set period, for example 15 months for a new hire. When a period ends, completed records stay.

Every module spends at least 70% of its estimated time on active work: case analysis, branching scenarios, worked exercises, simulations, knowledge checks, assessments and action plans the learner completes. Video, reading and reviewing model answers count as passive. The rule applies to each module on its own. It is never averaged across the catalogue.

Two choices keep the count conservative. Reviewing feedback and model answers counts as passive, even though it follows active work. And passive time is budgeted on the French text, which runs longer than the English, so the English version is at least as practical as the published figure.

In the specification the lowest module is 72.2% and the catalogue is 82.4%. The platform computes the ratio from the stored duration of each activity and will not publish a module that falls below 70%. Once learners are using it, every activity records its time, and administrators get a report of the measured practical share by module, from real learner durations rather than our estimates.

Assessment is progressive: knowledge checks inside each module, a scored assessment at the end of it, a capstone that certifies each level, and an annual currency check for senior bankers. Most scored items sit inside a case.

Item types include numeric entry for ratios and coverage, marking red flags on a financial exhibit, spreading by drag and drop (with a keyboard alternative), building a loan structure within an expert range, branching conversations, memo sections scored on a rubric, and a committee simulation that scores whether the learner can defend a position and revise it when the facts call for it.

Six critical errors fail a certification whatever the total score, because they are the ones that would hurt a real portfolio. Among them: approving below 1.0× coverage with no credible mitigant, relying on security that cannot be enforced as described, and missing a Crown priority claim for unremitted source deductions.

Feedback is specific to the answer given. A wrong answer that matches a known misconception gets feedback written for that misconception. After each scored assessment the learner sees their result by sub-competency against the level target, and one recommended next step.

Retakes and remediation. A failed attempt opens a second one after 24 hours, on a parallel form drawn from a pool at least three times the length of the test, so a retake never repeats the same questions. A second failure assigns short targeted activities before a third attempt. Every attempt stays in the learner's history.

Certification memos are scored by trained human assessors. Ten per cent are scored twice to keep assessors consistent.

Each module ends with a step that points to a tool the banker uses on real work: a discovery conversation guide, an information request checklist, a ten-minute scan for ASPE statements, a recommendation self-review. Managers get their own coaching guides, a file-review rubric and a checklist for joint client visits. The full list of fourteen tools is on the curriculum page. The tools are static question lists and prompts. They do not process client data.

Passing a level capstone earns a digital credential for that level. Learners can download it, share it (on LinkedIn, for example) and give anyone a link to verify it. Credentials can be revoked, and verification shows when they have been. The annual currency pathway issues a separate badge.

Many institutions already run a learning management system and want records in one place. Crediteris is designed to fit in three ways:

• SCORM packages. Modules export as SCORM 1.2 or SCORM 2004 packages that your LMS hosts and tracks. This is the main path for Workday Learning, which accepts SCORM and AICC but not xAPI. Crediteris does not support AICC.
• LTI 1.3 launch. Your LMS launches learners into Crediteris with single sign-on, and completions and scores flow back.
• xAPI records. Every activity is recorded as an xAPI statement, with timestamps, attempts and scores. Statements can be forwarded to your own learning record store.

A documented REST API and webhooks cover enrolment, progress and export. At the end of a contract every learning record exports in a standard format, so you can move to another system.

Each module has an English lead and a French lead, both credit professionals, and the French is written in Québec French rather than translated at the end. A shared termbase fixes the credit vocabulary against the French edition of the CPA Canada Handbook. Where the law differs, each language teaches the right concept: a movable hypothec published at the RDPRM is not a PPSA security agreement, and the course says so in both languages. A release ships only when both versions pass review. Learners can switch language at any time without losing progress.

Every part of Crediteris works without AI. An institution can turn on AI features one at a time: draft feedback on a practice memo, adaptive follow-up questions in the committee simulation drawn from a fixed question bank, and ordering of practice items. AI sees only the fictional case and the learner's answer, never client data. A human assessor confirms every certification score, and a learner can ask for human review of any AI feedback. We will name the model and every sub-processor behind each AI feature before it can be enabled.

Crediteris runs in current versions of Edge, Chrome and Safari on laptops and desktops, and on tablets in either orientation. There is no app to install.