UNVPortal Quality: collects the evidence the moment it is created

A quality management system is software that records an institution's processes, committee decisions, improvement actions and their evidence against accreditation criteria. The PDCA cycle (plan, do, check, act) is a four-stage loop that keeps an improvement action from closing until its result has been measured. UNVPortal Quality is a quality management system of this kind, written for the self-assessment and accreditation work of higher education institutions.

• Every member of staff: sees the task that falls to them, completes it within the deadline and uploads the evidence. Quality work stops being a pile of documents collected by a few people at the end of term and becomes a trace recorded as the work is done.
• Academic staff: tasks linked to course, programme and student outcomes; syllabi, assessment evidence, improvement notes.
• Administrative staff: process steps, service standards, records and evidence for satisfaction and support processes.
• Unit quality representative: monitors the task status of their unit, sees missing evidence, compiles the unit self-assessment report.
• Committee member: meeting agendas, decision records and the follow-up of actions arising from decisions.
• Quality coordination office: criterion tracking, evidence health, consolidation of unit reports into the institutional report.
• Senior management: open actions, overdue items and changes in indicators across terms.

Staff with no assigned task do not meet an empty screen: they see the open items that fall to their unit and the evidence expected of them.

Tasks and evidence

Work arising from a criterion falls to the responsible member of staff as a task. On their own screen they see what needs to be done, the deadline and which evidence is expected; when the work is done, the evidence is uploaded there. A task without evidence does not close and shows as missing in the unit report.

Because the whole institution uses the same screen, quality data does not depend on one unit's end-of-term effort: it is recorded as the work is done. Who completed which task and when remains in the audit log.

Process management

The institution's processes are defined and versioned with their owner, inputs and outputs; the flow is drawn on a canvas in the browser. The process catalogue shows who owns which process and how processes depend on one another. Running processes are tracked as instances.

On the performance dashboard, target attainment, the number of observations and data quality sit in separate columns: a problem with the measurement itself is not read as a drop in performance. Processes whose review date has come fall into a queue and are worked through as campaigns — a review date cannot pass unnoticed.

Committee management

Committees are set up with members, terms of reference and a meeting calendar; which committees a person sits on is visible on a single screen.

Meetings, decisions and actions

The meeting agenda is prepared, the decision is written into the minutes, and the action arising from the decision is opened with its owner and target date. Overdue actions are collected in a separate list. The next meeting's agenda starts with the open actions of the previous one.

Evidence management

Evidence is a record linked to a criterion. A document can be uploaded, or a link given to a record that already exists in the system: a committee decision, an instrument usage report, a programme outcome result. This stops the same document multiplying across folders.

The PDCA cycle

Every action passes through plan, do, check and act. An action whose measurement result has not been entered at the check stage does not close; it shows as open in the term report. Whether the improvement worked is thereby put on record.

Self-assessment reports

The report template is built from criterion headings. The evidence and indicators linked to each heading come from the system; text fields are filled in by the unit. Unit reports are consolidated into the institutional report, and the system lists which unit has left which heading empty.

Nonconformities and corrective action (CAPA)

A nonconformity arising from an audit, a complaint or an indicator missing its target is recorded with its source, severity, class and owner. Root cause analysis is carried out with a method; corrective and preventive actions are attached to it. Closure passes through a gate: a CAPA whose effectiveness has not been verified does not close.

Overdue CAPAs and those approaching their deadline are collected on a separate dashboard with the days remaining. Recurrence analysis shows how many times the same nonconformity has come up and in which process — a recurring nonconformity is evidence that a correction counted as closed did not work.

Risk and opportunity

Risks are assessed on the institution's own matrix and shown on a heat map; opportunities are kept in a separate list. A risk, too, passes through a gate as it closes.

Internal audit

The audit plan, the checklist template and auditor assignments are set up in the system. During field work notes are written, a nonconformity is turned into a finding and evidence is attached; the CAPA arising from a finding opens in the same chain. The package for an external assessor is compiled from the plan in one go.

Management review

The agenda is built from a template and its inputs come from the system: indicators, open CAPAs, audit findings, risks. Decisions taken are opened as actions and enter the same follow-up as every other action.

Indicator tracking

Indicators are defined with their target, unit and period; the trend is visible across periods. The PDCA closure check looks at whether the indicator actually moved after the improvement.

Obligation inventory

The regulatory provisions binding on the institution are recorded as obligations, with their owner, assessment schedule and the evidence expected. The matrix shows which obligation maps to which process and criterion; the readiness index says which obligation is left without evidence.

Regulatory change and impact analysis

When a watched source changes, a record is opened and matched to the relevant provisions. Before it is accepted, an impact preview is produced: which obligations, processes and criteria this change affects. On acceptance, impact records are generated and the owners are notified.

Acceptance is a separate permission, and holding that permission is not enough on its own: the person accepting must be someone other than the person who entered the change, and must own the affected records. One person cannot accept a change they entered themselves.

Standard versions and criterion mapping

Two versions of a standard are compared and the changed clauses are extracted; clause lineage shows a clause's counterpart in the previous version as a chain. A crosswalk is built between two sets of criteria, and gap analysis lists which criterion is left uncovered — this is how you see which evidence can be reused when moving from one accreditation to another.

A comparison changes no obligation by itself: the machine proposes, a person publishes.

Extracting data from minutes

When scanned meeting minutes are uploaded, decisions, attendees and actions are extracted field by field; each field's confidence level and its position in the document are shown. Where confidence is low, a person picks the match. Running the extraction, viewing it and importing it are separate permissions, and extraction finalises nothing — closing a meeting or formalising a decision belongs to another permission.

Report generation

The report is generated deterministically from evidence records: the same data produces the same output every time. Every claim is linked to a criterion clause and to at least one piece of evidence; a claim without evidence is marked unsupported and the report does not pass validation. "Not yet measured" and "measured and not met" are separate states, and the second one is not hidden.

AI can write report text, but it does not produce the report's facts: it turns the findings of the deterministic engine into sentences and adds them as draft claims. The provenance of the text can be viewed, nothing is published without human approval, and a claim without evidence does not pass approval in the first place.

Data room

No account is created for an external assessor on the institution's system. With an invitation link, a one-time code sent to their email and a time-limited session, they see only the sections and documents opened to them; every view and every download is logged. They ask questions — the answer is published once the institution approves it; they request additional evidence; they see the visit agenda. When the period ends or access is revoked, the session closes immediately.

Part of quality work does not happen at a desk. In the mobile app an auditor writes notes in the field, turns a nonconformity into a finding and attaches the evidence on the spot; a document is scanned with the phone camera and uploaded. Tasks, evidence, meetings and decision tracking are visible on the phone too.

With no connection the record is queued on the device and sent when the network returns. The app shows what has been sent separately from what is still waiting: marking both the same way would let an auditor believe they had done something they had not. If the record on the server changed in the meantime, the conflict is put to the user rather than silently overwritten.

Permissions and identity

Permissions are granted through a permission catalogue. The institution connects its own identity provider (SSO); user mapping and role rules are defined, and federation sessions and identity audit events are visible.

Competency and delegation

A task can carry a competency requirement: someone who does not meet it cannot take the work on, and that block is recorded. A warning is raised before competency certificates expire. During leave or an assignment, a permission is delegated to a deputy for a set period; out-of-the-ordinary access is logged separately and reviewed afterwards.

Connectors

To bring in data from the institution's other systems, SFTP file transfer and a read-only database view are connected; mapping is set up without writing code. Connection health, errors and records that could not be processed are monitored on their own screen.

Quality can be deployed on its own. Deployed together with Labs, instrument usage reports are linked directly as evidence. Deployed together with Exams, programme outcome results come from the system as report indicators instead of being calculated by hand.

It runs in the browser. Permissions are granted per unit and committee; a unit representative sees only their own unit's records, the quality coordination office sees everything. The institution can choose a cloud deployment or its own server.

Note: this is quality management software. It carries no authority to accredit, audit or certify on behalf of any body; it supports the institution in running and reporting its own processes.

Deployment and licensing

It is enough to contact us with your institution's name, the number of units and your reporting calendar; your enquiry is answered with proposed times for a live demo.