# Internal Quality Assurance for Credential Issuers *Quality between reviews: approval, monitoring and periodic review.* ## About this course **Course ID:** ICB-A05 **Framework alignment:** ICoBC Credentialing Skills Framework (ICSF), skill QUAL "Quality assurance and accreditation" (VRITE pillar: Rigor), level L4 "Assure", stage A "Ability". **Badge:** "Internal Quality Assurance for Credential Issuers (Ability, L4)", alignment code ICSF-QUAL-L4-A. ### Who this course is for This course is for people who run or support credential quality inside an issuing organisation: - programme and micro-credential leads - quality, registry and admissions staff - learning and development (L&D) and corporate academy leads - professional bodies and sector associations - ICoBC Lead Reviewers You should have completed ICB-P18 Preparing for Quality Review: the VRITE Self-Assessment. ### What you will be able to do At the Ability stage you practise on realistic tasks in a controlled setting. At level L4 "Assure", the skill is to design and lead internal quality assurance and to conduct independent reviews of other issuers. By the end of the course you will be able to: 1. Design an internal quality cycle for credentials: design approval, monitoring and review. 2. Define roles, responsibilities and independence within the issuer. 3. Use data and feedback to improve credentials. 4. Prepare for external review and maintain the Seal. ### Workload and format The course is blended and runs as a cohort. The notional learning time is 12 hours. This file covers the self-paced part. Live or workshop elements appear as facilitator notes. ### How it is assessed You complete a short knowledge check in this file. A person then assesses your final task with a rubric: an outline of an internal quality manual for an issuer, with process maps. The quality manual outline is the evidence for the badge. Completing the course does not prove competency in a job. The badge records that you met the stated criteria in a controlled setting. ### Outcomes, modules and assessment | Outcome | Module | How it is assessed | |---|---|---| | 1. Design an internal quality cycle for credentials | Quality systems | Module quiz; knowledge check; quality manual outline (cycle and process maps) | | 1. Design an internal quality cycle for credentials | Design approval | Module quiz; knowledge check; quality manual outline (approval process map) | | 2. Define roles, responsibilities and independence | Quality systems | Module quiz; knowledge check; quality manual outline (roles section) | | 2. Define roles, responsibilities and independence | Design approval | Module quiz; knowledge check; quality manual outline (approval panel and independence) | | 3. Use data and feedback to improve credentials | Monitoring | Module quiz; knowledge check; quality manual outline (monitoring plan) | | 3. Use data and feedback to improve credentials | Periodic review | Module quiz; knowledge check; quality manual outline (review and action tracking) | | 4. Prepare for external review and maintain the Seal | Periodic review | Knowledge check; quality manual outline (evidence base) | | 4. Prepare for external review and maintain the Seal | External review interfaces | Module quiz; knowledge check; quality manual outline (external review readiness) | ### AI focus Many issuers now use AI-supported processes, for example to draft learning outcomes, screen applications, mark assessments or summarise learner feedback. Each module includes a short note on quality checks for such processes. ### Status of reference topics | Topic | Status | Note | |---|---|---| | ICoBC VRITE criteria and Global Quality Seal | Emerging | Indicators are not yet published. Internal documents use different expansions of the acronym. Pilots are ongoing. | | EU Council Recommendation on micro-credentials (2022) | Established | Stable reference. Watch for Commission implementation reporting and national transpositions. | Status: November 2026. --- ## Quality systems ### Why issuers need an internal quality system External review looks at an issuer at a point in time. Learners, employers and partners rely on the credential every day in between. An internal quality system is the set of planned, documented and owned activities that keep credentials sound between reviews. A credential issuer carries a promise: this credential means what it says it means. The internal quality system checks that the promise holds. It does this by asking the same questions again and again: - Is the credential well designed before it is launched? - Does it work as intended while it runs? - Does it still fit its purpose after some time? - Can we show evidence for all of this? The ICoBC VRITE criteria group quality into five pillars. An internal system should touch all five. The ICoBC acronym is used in different expansions in internal documents, and the indicators are not yet published, so this course uses the pillars as headings for your own questions and does not rely on a fixed indicator list. Where your manual refers to the pillars, name them as your organisation or the ICoBC pilot documents define them. ### The quality cycle A cycle has three connected stages and one shared evidence base. 1. **Design approval:** a credential is checked before it is offered or changed. 2. **Monitoring:** data and feedback are collected and acted on while the credential runs. 3. **Periodic review:** at set intervals, the whole credential is re-examined and a decision is made to continue, change or withdraw it. The evidence base holds the records that make each stage visible: approval decisions, data reports, feedback summaries, action logs and review reports. ### Roles, responsibilities and independence A cycle only works if people own it. Define at least these roles, and write them down: - **Credential owner:** leads the credential and answers for its quality day to day. - **Quality lead:** runs the quality system, keeps the calendar and the records. - **Approval panel or review panel:** makes the approval and review decisions. - **Governing body or senior sponsor:** receives reports and resolves escalations. - **Assessors and issuers of the credential:** apply the criteria as written. Independence is the key design question. The person who designs a credential should not be the only person who approves it. A review panel should include at least one member who does not work on the credential. In small organisations, independence may come from another department, a partner organisation or an external expert. Record who is independent and from what. **Worked example: Roles in two organisations** *University, college or training provider.* A college offers a 10-credit micro-credential in data literacy. The programme lead is the credential owner. The registry manager acts as quality lead and keeps the approval and review calendar. The approval panel is the college's academic standards committee, chaired by a deputy head from a different faculty. Independence comes from the chair and from one external employer representative. The senior sponsor is the vice-principal for teaching. *Corporate academy or HR department.* A corporate academy issues a badge for "Safe handling of customer data". The academy lead is the credential owner. An HR business partner acts as quality lead. The approval panel includes a compliance officer from the legal team and a line manager from a business unit that uses the credential. Independence comes from the compliance officer, who has no role in training delivery. The senior sponsor is the chief people officer. ### AI-supported processes in the quality system If a process uses AI, add it to the quality system as a named process with an owner. Typical checks: - A human owner is named and accountable for the output. - The purpose of the AI use is written down. - Outputs are sampled and checked by a person against the criteria. - Learners and earners are told where AI is used in decisions that affect them. - Changes to the AI-supported process go through the same approval route as other changes. **Practice activity: Draw your quality cycle** Choose a real or realistic credential (your own, or a case supplied by your facilitator). Complete these steps in about 45 minutes: 1. Sketch the three stages on one page. Add the main inputs and outputs of each stage. 2. List the roles in your organisation that would fill the five roles above. If a role has no owner, say so. 3. Mark where independence comes from at each decision point. 4. Mark one AI-supported process, if any exists, and say who is accountable for it. 5. Write two sentences: what is the weakest point in your cycle, and what would you change first? Keep your sketch. You will reuse it in the final task.

Facilitator notes

Run a 90-minute workshop after this module. Put participants in small groups of mixed backgrounds (providers, companies, associations). Each group presents its cycle sketch in three minutes. Ask the others to find one independence gap. Collect common gaps on a flipchart for use in the Design approval module. Remind groups that numbers or names they use must be illustrative.

### Check your understanding **Question 1.** What is the main purpose of an internal quality system for a credential issuer? [( )] To replace external review so that no outside body needs to look at the issuer [(X)] To keep credentials sound between external reviews and to produce evidence that they are [( )] To make credentials harder to earn [( )] To record which staff members are responsible for marketing [[?]] Think about what happens between two external reviews. **** The internal system runs continuously and creates the evidence that supports any external review. It does not replace external review, and its aim is not to make credentials harder to earn. **** **Question 2.** Which three stages form the core internal quality cycle in this course? [( )] Marketing, enrolment and certification [( )] Planning, delivery and invoicing [(X)] Design approval, monitoring and periodic review [( )] Registration, accreditation and withdrawal [[?]] The stages cover before launch, while running and at set intervals. **** Design approval checks a credential before launch or change. Monitoring collects data and feedback while it runs. Periodic review re-examines the whole credential at set intervals. **** **Question 3.** A programme lead designs a micro-credential and is also the only person who approves it. What is the main quality risk? [( )] The credential will cost too much [( )] The approval will take too long [(X)] The approval lacks independence, so design weaknesses may go unnoticed [( )] The credential cannot be offered online [[?]] Consider who checks the designer's work. **** Independence means that someone other than the designer examines the design. Without it, weaknesses that the designer cannot see may remain. **** **Question 4.** An issuer uses an AI tool to summarise learner feedback. Which step belongs in its quality system? [( )] Switch off human review to save time [( )] Treat the summaries as final and archive the raw feedback [(X)] Name an accountable owner and have a person sample the summaries against the raw feedback [( )] Keep the use of AI secret from learners [[?]] AI-supported processes still need human accountability. **** A named owner and regular human sampling keep the process under control. Removing human review or hiding the use of AI weakens quality and trust. **** --- ## Design approval ### What design approval is for Design approval is the gate before a credential is offered or materially changed. It answers one question: is this credential well designed and ready to run? A good gate is proportionate. A new credential gets a full panel review. A small change, such as an updated reading list, may only need a documented check by the credential owner and the quality lead. Write the approval route down. People should know in advance what evidence they must bring, who decides and how long it takes. ### What the panel looks at A design approval pack lets the panel judge the credential against criteria. Typical content: - **Purpose and need:** who the credential is for and why it is needed. - **Learning outcomes:** clear statements of what an earner can do, at a stated level. - **Assessment:** how achievement is judged, and how the assessment matches the outcomes. - **Delivery and workload:** how learners will learn, and the notional workload. - **Resources and staff:** whether people and materials are sufficient and competent. - **Recognition and information:** what the credential says about itself, including the information that the EU Council Recommendation on micro-credentials (2022) lists as elements of a micro-credential description. - **Fairness and access:** how learners with different needs and backgrounds are supported. - **Risks:** including data protection and the use of AI. The EU Council Recommendation on micro-credentials (2022) is a stable reference for the information that should accompany a micro-credential. Check the current text for the exact list of elements and use it as one source for your pack template. Watch for Commission implementation reporting and national transpositions, which may add requirements in your country. ### Decisions and conditions The panel should reach one of four decisions: 1. **Approved.** 2. **Approved with conditions:** with named actions, owners and deadlines. 3. **Deferred:** more evidence is needed before a decision. 4. **Not approved:** with written reasons. Record the decision, the people who took part, any declared conflicts of interest and the conditions. Track conditions until they are closed. An approval with open conditions that nobody checks is a common weakness. **Worked example: Approving a new credential** *University, college or training provider.* A training provider proposes a 5-credit micro-credential on sustainable procurement. The credential owner submits the pack. The quality lead checks completeness and finds that the assessment section does not say how the final project is marked. The pack is returned within three working days (illustrative). After correction, the panel meets. The independent chair, a senior lecturer from another department, asks how the project marks map to the three learning outcomes. The panel approves with one condition: publish the marking criteria to learners before the first run, owner to confirm to the quality lead within one month (illustrative). *Corporate academy or HR department.* An academy proposes a badge for "Inclusive recruitment interviewing". The pack states that managers will use an AI tool to draft interview questions. The compliance officer on the panel asks who checks the drafted questions for bias. The academy lead had not planned a check. The panel defers the decision. The owner adds a step in which two trained interviewers review a sample of generated questions each quarter, and the pack returns for approval. ### Quality checks for AI-supported design AI tools can help draft outcomes, assessment items and course text. Add these checks to the approval route: - Declare in the pack where AI was used in design. - A subject expert checks every outcome and assessment item against the purpose. - A person checks that sources and facts in AI-drafted material exist and are correct. - Check for bias in language and examples. **Practice activity: Review an approval pack** Your facilitator supplies a short pack for a fictional credential, or you can use a pack from your own organisation with identifying details removed. Complete these steps in about 60 minutes: 1. Read the pack and list the evidence that is present and the evidence that is missing. 2. Write three questions that the panel should ask the credential owner. 3. Choose a decision: approved, approved with conditions, deferred or not approved. Write the decision and the reasons in three to five sentences. 4. If you chose conditions, write each condition with an owner and a deadline (illustrative). 5. Note one declared conflict of interest that the panel might need to handle, and say what it should do. Keep your approval route and decision template. You will reuse them in the final task.

Facilitator notes

Run a 2-hour simulated panel. Assign roles: credential owner, quality lead, independent chair, two panel members, and one observer who records decisions. Give each panel a different fictional pack, one of which contains a deliberate gap in assessment design and one of which contains undeclared AI use. After the panel, debrief on how independence affected the questions asked. Ask observers to check that the decision record contains reasons, conditions, owners and deadlines.

### Check your understanding **Question 1.** What is the main question that design approval answers? [( )] Is the credential popular enough to recruit learners? [( )] Is the credential cheaper than competitors? [(X)] Is the credential well designed and ready to run? [( )] Has the credential been marketed to employers? [[?]] Design approval is a gate before launch or change. **** Design approval checks design quality and readiness. Popularity, price and marketing are separate concerns. **** **Question 2.** A panel approves a credential but lists three conditions with no owners or deadlines. What is the main weakness? [( )] Panels should never approve with conditions [(X)] Without owners, deadlines and tracking, conditions may never be closed [( )] Conditions should only be set by the credential owner [( )] The panel should have deferred every time [[?]] Think about who will act on each condition. **** Conditions need an owner, a deadline and tracking until closed. Otherwise the approval hides open risks. **** **Question 3.** A small change, such as an updated reading list, is proposed for an approved credential. What is the most proportionate approval route? [( )] A full panel review with external experts [( )] No record, because the change is small [(X)] A documented check by the credential owner and the quality lead [( )] Withdrawing the credential until the next periodic review [[?]] Proportionate means effort matches risk, and still leaves a record. **** Small changes can use a lighter route, but the decision and the reason should still be recorded. **** **Question 4.** A design pack states that AI helped to draft assessment items. Which check should the approval route require? [( )] None, because AI output is always accurate [( )] Only a spelling check [(X)] A subject expert checks the items against the learning outcomes and the sources [( )] The AI tool vendor signs off the items [[?]] Someone with subject knowledge must stay accountable. **** A subject expert checks alignment with outcomes and confirms facts and sources. AI output can contain errors, and accountability stays with people. **** --- ## Monitoring ### What monitoring is for Monitoring starts after launch. It collects data and feedback while the credential runs, so that problems are found early and good practice is spotted. Monitoring is routine and light. Periodic review, which comes next, is deeper and less frequent. A monitoring plan answers five questions: 1. What do we want to know? 2. Which data or feedback will tell us? 3. Who collects it, and how often? 4. Who looks at it, and when? 5. What happens when a signal crosses a threshold? ### Choosing indicators Choose a small set of indicators that link back to the purpose of the credential. Too many indicators create noise. Examples: - **Participation and completion:** starts, completions and withdrawals. - **Assessment outcomes:** distribution of results, appeals, and consistency between assessors. - **Learner feedback:** satisfaction and comments about clarity, workload and fairness. - **Earner and employer feedback:** whether the credential is understood and useful in practice. - **Access and fairness:** whether results and completion differ between groups of learners in ways that need explanation. - **Integrity:** cases of suspected misconduct, and the handling of the credential record. - **Issuing process:** time to issue, errors in credential records, and requests for correction. Use thresholds and triggers, set in advance. For example, a completion rate below an agreed level triggers a short analysis by the credential owner. Numbers in this course are illustrative. Your organisation sets its own thresholds with its governing body. Respect data protection rules when you collect and analyse learner data. Collect only what you need, tell learners how data is used, and store it securely. ### From data to improvement Data only helps if it leads to action. Use a simple action log with these columns: finding, evidence, action, owner, due date, status and effect. Review the log at a regular quality meeting. Close an action only when the effect has been checked. Take care with interpretation. A low completion rate may mean the credential is too hard, too long, badly timed or poorly explained. Combine numbers with feedback and, where possible, with a conversation with learners and assessors before deciding what to change. **Worked example: Reading monitoring data** *University, college or training provider.* A college's micro-credential in project management shows a drop in completion from 82 per cent to 64 per cent between two cohorts (illustrative figures). The threshold for analysis is 70 per cent (illustrative). The credential owner reviews feedback and finds that many learners mention a heavy workload in week four. Assessors confirm that the week-four task takes longer than the notional hours suggest. The action is to split the task into two shorter tasks and to update the workload statement. The owner logs the action, the due date and the check: completion rate for the next cohort. *Corporate academy or HR department.* An academy issues a badge for "Customer data protection basics". Quarterly data show that the badge is earned quickly by most employees, but a line-manager survey reports that the same errors continue in practice. The academy lead concludes that the assessment tests recall, not application. The action is to add a scenario-based task, and to ask two line managers to review the scenarios. The badge description is updated to state clearly what the assessment covers. The badge states what was assessed and does not claim more. ### Quality checks for AI-supported monitoring If you use AI to analyse feedback or flag unusual results, apply these checks: - Compare AI summaries with a sample of the original comments each cycle. - Record the version and settings of the tool, so results can be repeated. - Check whether the tool treats groups of learners differently. - Keep a person responsible for every action that follows from an AI-generated flag. **Practice activity: Build a monitoring plan** Take the credential from earlier modules. Complete these steps in about 60 minutes: 1. Choose five indicators that link to the purpose of the credential. 2. For each indicator, write the source, who collects it, how often, and one threshold (illustrative). 3. For two thresholds, write the action that a breach triggers and who owns it. 4. Draw a one-page process map of the monitoring loop for your organisation. 5. Create an action log template with the seven columns named above, and fill in one example row. 6. Write a short note on one data protection point that applies to your indicators. Keep your monitoring plan and process map. You will reuse them in the final task.

Facilitator notes

Provide a small, fictional data set for one credential over three cohorts (completion, assessment results, five learner comments per cohort, one employer comment). Ask pairs to identify two signals and one possible misreading of the data. Run a 30-minute plenary on how to combine numbers with feedback. Emphasise that thresholds are set by the organisation and that figures in the exercise are illustrative.

### Check your understanding **Question 1.** What is the role of thresholds in a monitoring plan? [( )] They decide which learners may graduate [(X)] They set in advance the level at which a signal triggers analysis or action [( )] They replace periodic review [( )] They measure staff performance only [[?]] Thresholds are agreed before the data arrives. **** Thresholds and triggers, set in advance, tell the team when to look more closely and who acts. They do not replace periodic review. **** **Question 2.** Completion falls between two cohorts. What is the best first step? [( )] Lower the pass mark immediately [( )] Withdraw the credential [(X)] Combine the numbers with feedback from learners and assessors to understand the cause [( )] Ignore the change until the next periodic review [[?]] A number alone does not tell you why. **** A fall in completion has several possible causes. Feedback from learners and assessors helps to identify the right action. **** **Question 3.** When should an action in the action log be closed? [( )] When the owner says the action is done [( )] When the quality meeting ends [(X)] When the effect of the action has been checked [( )] When the next cohort starts [[?]] Closing an action should rest on evidence. **** An action is closed when its effect has been checked, for example through the indicator that triggered it. **** **Question 4.** An AI tool flags a group of learners as "at risk". What is the right quality check? [( )] Act on the flag automatically, without review [( )] Delete the flag to protect privacy [(X)] Have a named person review the flag and its basis before any action [( )] Share the flag with all staff [[?]] AI flags are inputs to human decisions. **** A named person reviews the basis of the flag before acting. This keeps accountability with people and protects fair treatment. **** --- ## Periodic review ### What periodic review is for Periodic review is a planned, deeper look at a credential, at a set interval. Monitoring keeps an eye on the credential from week to week. Periodic review steps back and asks whether the credential is still needed, still well designed and still fit for purpose. It ends in a decision: continue, change or withdraw. Set the interval in your quality manual. Intervals depend on risk and on the nature of the credential, for example, shorter for a credential in a fast-changing field. State the interval and the triggers for an earlier review, such as a serious complaint, a change in regulation or a major change in the field. ### What the review covers A review uses the evidence collected since approval, and asks fresh questions: - **Purpose and need:** is the need still there? Who uses the credential? - **Learning outcomes and level:** do they still match the field and the intended level? - **Assessment:** is it still valid, fair and consistent? - **Data and feedback:** what do the monitoring records show over the whole period? - **Earner and employer views:** is the credential understood and used? - **Resources and staff:** are they still adequate? - **Information and recognition:** is the description accurate, and does it still match the relevant references, such as the EU Council Recommendation on micro-credentials (2022)? - **Risks and incidents:** what has happened, and what has been learned? - **Previous actions:** were they completed and did they work? A review is more credible when someone independent leads or joins it. Include at least one panel member from outside the credential team, and consider an external expert. Reviews of other issuers, which an L4 practitioner may conduct, follow the same logic: use criteria, gather evidence, give findings and recommend actions. ### Withdrawal and teach-out If a review decides to withdraw a credential, plan the teach-out. Learners who have started must be able to finish or be offered a fair alternative. Earners who already hold the credential keep a valid record. Tell stakeholders in good time and keep the evidence. ### Evidence for external review A periodic review is a rich source of evidence for external review. Keep the review report, the panel membership, the data used, the decision and the action plan in a place where they can be found. This evidence shows that the issuer examines itself honestly and acts on what it finds. The ICoBC VRITE criteria and the Global Quality Seal are emerging: indicators are not yet published and pilots are ongoing. Design your evidence store so that it can be mapped to new indicators when they appear, instead of waiting for them. Status: November 2026. **Worked example: A periodic review decision** *University, college or training provider.* A university department reviews its micro-credential in digital accessibility after three years. The panel includes a lecturer from another faculty, an accessibility practitioner from a partner organisation and a student representative. The panel reads three years of monitoring records and sees stable completion and positive feedback. It also finds that two learning outcomes refer to a standard that has been revised. The decision is to continue with changes: the department updates the outcomes and the assessment, and the changes go through a light approval route. The next review is set in three years (illustrative). *Corporate academy or HR department.* An academy reviews a badge in "Workplace first aid awareness" after two years. Earner data show good completion. The review finds that the badge is not recognised by two business units, who run their own training. A manager from one unit joins the panel. The panel decides to change the credential so that it covers the content those units need, and to withdraw a duplicate internal course. A teach-out plan lets staff already enrolled in the duplicate course move to the badge route. The review report is stored with the monitoring records. ### Quality checks for AI-supported review When AI helps to prepare review documents or to analyse large sets of feedback, the panel should: - Read a sample of the underlying material, not only the AI summary. - Ask the tool's owner to explain what was done and with what settings. - Record in the review report where AI was used. - Keep the final judgement and the decision with the panel. **Practice activity: Plan and run a mini review** Use the credential from earlier modules, with a fictional three-year history supplied by your facilitator. Complete these steps in about 75 minutes: 1. Write a one-page review plan: interval, panel membership with independence, questions and evidence needed. 2. Read the supplied monitoring records and note three findings. 3. Draft a review decision: continue, change or withdraw, with reasons in five to eight sentences. 4. Write an action plan with at least three actions, each with an owner and a due date (illustrative). 5. Draw a process map of your periodic review, including the route for change and for withdrawal. Keep your review plan, decision template and process map. You will reuse them in the final task.

Facilitator notes

Run a 2-hour live session in which small groups act as review panels. Supply a fictional dossier: approval record, three years of monitoring summaries, two complaints and one change in the sector. Ask each panel to reach a decision in 45 minutes and present it in three minutes. Observers check that the independent member contributed and that the decision links to evidence. Close with a short discussion of teach-out duties and how to communicate a withdrawal.

### Check your understanding **Question 1.** How does periodic review differ in purpose from monitoring? [( )] Periodic review collects weekly data, monitoring does not [(X)] Periodic review is a deeper, planned re-examination that ends in a decision to continue, change or withdraw [( )] Periodic review is only needed after a complaint [( )] Periodic review replaces design approval [[?]] One looks at running data, the other takes a step back. **** Monitoring is routine and light. Periodic review is deeper and ends with a decision about the credential's future. **** **Question 2.** Which event might justify an earlier review than the planned interval? [( )] A new logo for the organisation [( )] A change of room for the exam [(X)] A major change in the field or a serious complaint [( )] A new staff member in the registry [[?]] Think about changes that affect the validity of the credential. **** Serious complaints, regulatory changes and major changes in the field are typical triggers for an earlier review. **** **Question 3.** A review panel decides to withdraw a credential. What must the issuer plan? [( )] Nothing, because the credential no longer exists [( )] Removal of all records of past earners [(X)] A teach-out so that enrolled learners can finish or move to a fair alternative, and valid records are kept [( )] A new credential with the same name [[?]] Consider learners who have already started. **** A teach-out protects learners already enrolled, and existing earners keep valid records. **** **Question 4.** Why is it useful to keep review reports in an organised evidence store? [( )] Because external reviewers read only the review report [(X)] Because the evidence shows that the issuer examines itself and acts, and it can be mapped to new indicators when they are published [( )] Because the store replaces the need for monitoring [( )] Because regulations require identical storage in every country [[?]] Evidence should be easy to find and reuse. **** An organised evidence store supports external review and can be mapped to indicators that are published later. It does not replace monitoring. **** --- ## External review interfaces ### What external review adds Internal quality assurance is the issuer's own work. External review adds an independent view. The reviewer checks whether the issuer's claims are true and whether its system works. A good internal system makes external review shorter, calmer and more useful, because evidence is ready and the issuer already knows its strengths and gaps. The ICoBC VRITE criteria and the Global Quality Seal are emerging. Indicators are not yet published, internal documents use different expansions of the acronym, and pilots are ongoing. For this reason, build your interface with external review around principles that will hold whatever the final indicators say: - You can show a documented cycle of approval, monitoring and review. - You can show who is accountable and who is independent. - You can show evidence for each claim you make about a credential. - You can show that findings lead to action, and that actions are checked. Status: November 2026. ### Preparing for external review Preparation is a routine, not a rush. Use this sequence: 1. **Scope:** agree which credentials and which processes will be reviewed. 2. **Self-assessment:** use the VRITE self-assessment from ICB-P18 to identify strengths and gaps. 3. **Evidence map:** list each claim and where its evidence sits. Fill gaps before the visit, or state them openly. 4. **Roles for the review:** name a lead contact, a records contact and a person who briefs staff. 5. **Briefing:** tell staff, learners and partners what to expect and what honesty and courtesy look like in a review. 6. **Logistics:** access to records, rooms or online sessions, and interpreters where needed. ### During and after the review During the review, answer questions directly and show evidence. If you do not know, say so and promise a follow-up. Do not coach staff to give rehearsed answers. After the review, treat the report as an input to your own cycle. Log each finding in the action log with an owner, a due date and a way to check the effect. Report progress to your governing body. If the reviewer recommends conditions, track them as you would approval conditions. ### Maintaining the Seal A seal or similar recognition is a promise that must be kept between reviews. To maintain it: - Keep the internal cycle running on its calendar. - Report significant changes to the credential or to the organisation to the relevant body, as its rules require. - Keep evidence up to date and findable. - Respond to findings within agreed times. - Watch for changes in criteria, indicators and procedures. For micro-credentials in Europe, the EU Council Recommendation on micro-credentials (2022) is a stable reference. Keep an eye on Commission implementation reporting and on national transpositions in the countries where you issue, because they may add requirements for information and quality assurance. Status: November 2026. ### Independent review of other issuers At level L4, you may also review other issuers. The skills are the same as internal review, turned outward: agree scope and criteria, gather evidence from documents and interviews, test claims, write findings that are clear and fair, and avoid conflicts of interest. Declare any past or present link with the issuer. Keep information confidential. Base every finding on evidence. **Worked example: Getting ready and following up** *University, college or training provider.* A college prepares for a review of its micro-credential portfolio. The quality lead builds an evidence map in a simple table: claim, source document, owner, date, status. The map shows that two credentials lack a recorded independent member in their latest approval. The quality lead states this openly in the self-assessment and sets two actions. After the review, the report recommends clearer records of conflicts of interest. The college adds a standing item to every panel agenda, logs the action with the registry manager as owner and checks the effect at the next three panels. *Corporate academy or HR department.* An academy prepares for an external review of its badge programme. The academy lead briefs line managers who employ earners, so that they can speak about how the badges are used. The evidence map shows gaps in the records of AI-supported marking. The academy documents the process, names an accountable owner and adds a quarterly sampling check. After the review, the findings go to the chief people officer, who agrees an annual progress report to the governing body. ### Quality checks for AI-supported evidence work If AI helps to prepare evidence maps, summaries or responses for external review: - Check every statement against the source record before it is submitted. - Do not submit text that you cannot explain or support. - Tell the reviewer where AI helped, if the review rules ask for it. - Keep a person accountable for the accuracy of the submission. **Practice activity: Evidence map and review preparation** Use your credential and your earlier work. Complete these steps in about 60 minutes: 1. Write five claims that your organisation would make about the credential, for example "assessment is independently moderated". 2. For each claim, name the evidence, where it is stored and who owns it. 3. Mark any claim for which evidence is weak or missing, and write an action for each gap. 4. Draft a one-page preparation plan for an external review, with named roles and a timeline (illustrative). 5. Write a short briefing note (about 150 words) for staff on what to expect and how to take part honestly. 6. Draw a process map from "review report received" to "actions closed". Keep your evidence map, plan and process map. You will reuse them in the final task.

Facilitator notes

Run a 90-minute role play. One group acts as the issuer, another as the external reviewers. The reviewers receive the issuer's evidence map and prepare five questions. The issuer answers using the map. Observers note where evidence was easy to find and where it was not. In the debrief, ask participants what they would change in their evidence store. Remind the group that the VRITE indicators are not yet published and that the exercise uses principles, not final criteria.

### Check your understanding **Question 1.** Why does a good internal quality system help with external review? [( )] It lets the issuer avoid answering questions [(X)] It makes evidence ready and shows the issuer already knows its strengths and gaps [( )] It allows the issuer to choose the criteria [( )] It removes the need for independence [[?]] Consider what the reviewer needs to see. **** Ready evidence and honest self-knowledge make review shorter and more useful. The reviewer, not the issuer, applies the criteria. **** **Question 2.** The VRITE indicators are not yet published. What is the sensible way to prepare? [( )] Wait until the indicators appear before doing anything [( )] Invent indicators and treat them as official [(X)] Build an evidence base around lasting principles: a documented cycle, accountability, independence and action on findings [( )] Stop internal monitoring until the pilots end [[?]] Prepare in a way that can be mapped to later indicators. **** Lasting principles can be mapped to indicators when they are published. Waiting or inventing indicators is not sound practice. **** **Question 3.** After an external review, how should the report be treated? [( )] Filed without action unless the reviewer asks again [( )] Shared only with the reviewer [(X)] Logged as findings in the action log, with owners, due dates and checks of effect [( )] Replaced by a new self-assessment [[?]] The report feeds the internal cycle. **** Findings become actions in the issuer's own system, with owners and checks, and progress is reported to the governing body. **** **Question 4.** When preparing evidence for a review with AI help, which practice is sound? [( )] Submit AI summaries without checking sources [( )] Hide the use of AI in every case [(X)] Check every statement against the source record and keep a person accountable for accuracy [( )] Let the AI tool answer the reviewers' questions [[?]] You must be able to explain and support what you submit. **** Every statement must be checked against the record, and a person remains accountable. The issuer should be open about AI use where review rules ask for it. **** --- ## Assessment and badge ### The assessed task **Task: Outline of an internal quality manual with process maps** You write the outline of an internal quality manual for a credential issuer. Use a real organisation (with permission and with sensitive details removed) or a realistic fictional issuer. The outline shows how the issuer designs and runs the internal quality cycle, who is accountable, how data and feedback lead to improvement and how the issuer prepares for external review. **Required content** 1. **Scope and purpose:** the issuer, the credentials covered, and the intended users of the manual (about half a page). 2. **Quality cycle overview:** one process map of the full cycle: design approval, monitoring, periodic review, and the links between them. 3. **Roles, responsibilities and independence:** a roles chart and a table of who decides what, with independence clearly marked. 4. **Design approval:** a process map, an outline of the approval pack, decision categories and condition tracking. 5. **Monitoring:** a monitoring plan with indicators, sources, frequency, illustrative thresholds, data protection note and the action log template. 6. **Periodic review:** interval and triggers, a process map covering continue, change and withdraw, panel composition and teach-out principles. 7. **External review interfaces:** evidence map, preparation steps, handling of findings and how the Seal would be maintained, based on principles and without assuming unpublished indicators. 8. **AI-supported processes:** a list of AI-supported processes in the issuer, with owner and quality checks for each. 9. **Document control:** version, owner, review date of the manual itself. **Format and length:** an outline of about 6 to 10 pages, including at least three process maps. Headings, tables and diagrams are welcome. Use illustrative figures only where labelled as such. Do not include personal data of real learners. **Evidence submitted for the badge:** the quality manual outline, as one document, with process maps included. Submit through the Kursbibliothek upload as instructed by your facilitator. A person assesses the task with the rubric below. The file does not grade the task. ### Rubric Each criterion is rated on four levels: 1 Not yet, 2 Developing, 3 Secure, 4 Strong. | Criterion | 1 Not yet | 2 Developing | 3 Secure | 4 Strong | |---|---|---|---|---| | A. Quality cycle design (Outcome 1) | Cycle missing or stages unconnected | Three stages named, links and outputs unclear | Three stages with clear inputs, outputs, links and calendar | Cycle is proportionate, includes triggers and routes for change and withdrawal, and fits the issuer's size and risks | | B. Roles, responsibilities and independence (Outcome 2) | Roles missing or vague | Roles listed, independence not addressed | Roles, decision rights and independence defined for approval and review | Independence is justified for each decision, conflicts of interest are handled, and gaps in small teams are addressed with workable solutions | | C. Process maps | No usable maps | Maps present but unclear, unlabelled or inconsistent with text | At least three clear, labelled maps consistent with the text | Maps show owners, decisions, records and feedback loops, and a reader could follow them without help | | D. Use of data and feedback (Outcome 3) | No monitoring plan | Indicators listed without sources, thresholds or actions | Indicators linked to purpose, with sources, frequency, thresholds and an action log | Indicators are few and well chosen, data is interpreted with feedback, actions are tracked to checked effect, and data protection is addressed | | E. Periodic review and improvement (Outcome 3) | Review not defined | Review mentioned, decision routes unclear | Interval, triggers, panel, decision options and action plan defined | Review links monitoring evidence to decisions, includes teach-out principles and shows how findings return to design approval | | F. External review and Seal maintenance (Outcome 4) | External review not addressed | Generic statements only | Evidence map, preparation steps and follow-up of findings defined | Preparation is a routine, gaps are stated openly, the approach holds whatever final indicators say, and Seal maintenance duties are handled with checks against official sources | | G. AI-supported processes | AI not addressed | AI named without owners or checks | AI processes listed with owners and checks | Checks are proportionate, include human sampling, transparency and change control, and are built into the cycle | | H. Clarity and accuracy | Frequent errors or unclear text | Understandable with gaps or errors | Clear, accurate, plain English, claims supported | Concise, well organised, accurate, no unsupported claims, illustrative figures labelled, and no overclaiming | ### Pass rule To earn the badge "Internal Quality Assurance for Credential Issuers (Ability, L4)" you need: 1. a score of at least 3 (Secure) on criteria A, B, D and F, and at least 2 (Developing) on every other criterion; 2. a score of at least 8 out of 10 on the knowledge check below; 3. a submitted quality manual outline that is your own work. If you used AI tools, say where and how, and confirm that you checked the content. If you do not pass, you receive written feedback and may resubmit once after revision. Your facilitator confirms timing. The badge records that you met these criteria in a controlled setting. It does not prove competency in a job. ### Knowledge check Answer all questions. Each question has one correct answer. **Question 1.** Which description best fits the internal quality cycle taught in this course? [( )] A yearly audit performed only by an external body [(X)] Design approval, monitoring and periodic review, connected and supported by an evidence base [( )] A single approval at launch, with no later checks [( )] A feedback survey sent once at the end of each cohort [[?]] Look for the answer with all three stages and the evidence base. **** The cycle has three connected stages and a shared evidence base. A single event or a lone survey is not a cycle. **** **Question 2.** An issuer wants the approval of a new credential to be independent. Which design best achieves this? [( )] The credential owner approves alone, to save time [( )] The programme team votes and the majority decides [(X)] A panel that includes at least one member who does not work on the credential [( )] The marketing team approves the credential [[?]] Independence means someone other than the designer examines the design. **** A panel with at least one member from outside the credential team provides an independent view. Self-approval and team votes do not. **** **Question 3.** In a very small organisation, no staff member is independent of a credential. What is a workable solution? [( )] Skip approval and rely on monitoring [(X)] Draw independence from another department, a partner organisation or an external expert, and record it [( )] Ask the credential owner to promise to be objective [( )] Delay the credential indefinitely [[?]] Independence can come from outside the credential team or the organisation. **** Independence can come from another department, a partner or an external expert. The source should be recorded. **** **Question 4.** A design approval panel decides "approved with conditions". What must the record contain? [( )] Only the word "approved" [( )] The names of learners who may enrol [(X)] The decision, reasons, panel members, declared conflicts of interest, and conditions with owners and deadlines [( )] The marketing plan [[?]] A decision record should let someone else follow what happened. **** A complete record includes the decision, reasons, participants, declared conflicts and tracked conditions with owners and deadlines. **** **Question 5.** Completion on a credential falls below the agreed threshold. What is the best next step? [( )] Lower the standard of assessment at once [( )] Wait for the next periodic review [(X)] Analyse the data with learner and assessor feedback, agree an action with an owner and log it [( )] Remove the data from the report [[?]] Use thresholds as triggers for analysis and action. **** Thresholds trigger analysis. Combining data with feedback leads to a justified action, which is logged with an owner. **** **Question 6.** An organisation uses an AI tool to summarise learner comments. Which quality check is appropriate? [( )] None, because the tool is automated [( )] Publish the summaries without review [(X)] A named owner samples the summaries against the original comments and records the tool's settings [( )] Delete the original comments [[?]] Human accountability and sampling keep AI support under control. **** A named owner, human sampling and records of settings keep the process accountable and repeatable. **** **Question 7.** A periodic review finds that a credential duplicates another course and no longer meets a need. Which decision path is correct? [( )] Continue without comment [( )] Delete all records, including those of earners [(X)] Withdraw or change the credential with a documented decision and a teach-out plan for enrolled learners [( )] Wait for a complaint before acting [[?]] The review ends in a decision, and learners must be treated fairly. **** The review decision should be documented. If it withdraws the credential, a teach-out protects enrolled learners and records of earners remain valid. **** **Question 8.** Why is a reviewer's report best treated as an input to the internal cycle? [( )] Because reports have no value after the visit [(X)] Because findings become logged actions with owners, due dates and checks of effect, which feed monitoring and review [( )] Because the report replaces the need for periodic review [( )] Because only the reviewer should act on it [[?]] Think about how findings lead to improvement. **** Treating findings as logged actions links external review to the internal cycle and drives improvement. **** **Question 9.** The VRITE indicators are not yet published. What approach to external review readiness is sound? [( )] Treat any draft you find as the official criteria [( )] Postpone all evidence work until the pilots end [(X)] Keep a documented cycle, clear accountability, independence and an evidence map that can be mapped to indicators later [( )] Rely on memory instead of records [[?]] Prepare with principles that hold whatever the final indicators say. **** Sound preparation uses lasting principles and an evidence base that can be mapped to the indicators when they are published. **** **Question 10.** Which statement about the EU Council Recommendation on micro-credentials (2022) fits the status given in this course? [( )] It is an emerging text whose content is still under draft [(X)] It is a stable reference, and issuers should watch for Commission implementation reporting and national transpositions [( )] It has no relevance to quality assurance of micro-credentials [( )] It replaces the need for internal quality assurance [[?]] The status given in this course is "Established". **** The Recommendation is a stable reference. Implementation reporting and national transpositions may add requirements, so issuers should keep watching them. Status: November 2026. **** --- ## Sources and further reading Status: November 2026. - Council Recommendation of 16 June 2022 on a European approach to micro-credentials for lifelong learning and employability, Council of the European Union, published in the Official Journal of the European Union. Check the Official Journal reference and the current text on the EUR-Lex service before you rely on it.- ICoBC VRITE criteria and Global Quality Seal: publications and pilot documents by ICoBC (International Council on Badges and Credentials, Berlin). Indicators are not yet published, and internal documents use different expansions of the acronym.- ICoBC Credentialing Skills Framework (ICSF): framework documents by ICoBC, including skill QUAL "Quality assurance and accreditation".- ICB-P18 Preparing for Quality Review: the VRITE Self-Assessment, ICoBC course (prerequisite for this course). - ISO 9001, Quality management systems: Requirements, International Organization for Standardization. A general reference on quality management systems. Check the current edition.- Standards and Guidelines for Quality Assurance in the European Higher Education Area (ESG), European Association for Quality Assurance in Higher Education (ENQA), together with the European Students' Union (ESU), the European University Association (EUA) and the European Association of Institutions in Higher Education (EURASHE). Check the current version. Do not rely on any single source for legal obligations. Check national laws and regulations that apply to your organisation and to the countries where you issue credentials.