What should you define before measuring PQRS?
PQRS is a Colombian acronym for petitions, complaints, claims, and suggestions. You'll also see PQRSD, which includes reports of alleged wrongdoing.
First, define the reporting period and cutoff date. A received case belongs to the period when it arrived. A closed case belongs to the period when it received a final response.
A final response should address the request and meet your organization's closure criteria. An acknowledgment confirms receipt, but doesn't show that the case was resolved.
For pending cases, take a snapshot at the cutoff date. Separate cases still within their deadlines from those overdue without a final response.
If a case is closed and later reopened, keep both dates and define how it appears in each report. To reconcile the balance, add cases open at the start and cases received. Subtract closures and account for reopenings.
In Colombia, check the deadline for the request type and your organization's activities. Colombia's Law 1755 of 2015 distinguishes between types of petitions and allows special rules. This guide isn't legal advice.
Which PQRS metrics can you use?
These formulas are a starting point. Document their dates, populations, and criteria before comparing results.
| Metric | Formula | What it shows |
|---|---|---|
| Received | Count of cases received during the period | Incoming workload |
| Closed | Count of cases with a final response during the period | Completed cases, including those received earlier |
| Timeliness | Cases closed within the deadline ÷ cases closed × 100 | Share of closures completed on time |
| Overdue open cases | Overdue open cases without a final response ÷ open cases × 100 | Share of open cases past their deadlines |
| Quality | Reviewed responses meeting the criteria ÷ reviewed responses × 100 | Quality of the reviewed sample |
| Satisfaction | Satisfied people ÷ survey respondents × 100 | Opinion of those who responded |
Calculate timeliness and quality using closures or reviews from the period. Calculate overdue cases using open cases at the cutoff date. If the denominator is zero, report “not applicable.”
How do you apply the formulas?
This example is illustrative. It doesn't describe a client or set a target.
A company receives 100 PQRS cases during the month. For simplicity, all the figures below refer to those same cases. At the cutoff date, 80 have a final response and 20 remain open.
Of the 80 closed cases, 72 received a response within the deadline. Timeliness is 72 ÷ 80 × 100: 90%. That result excludes open cases.
Of the 20 open cases, 5 are overdue without a final response and 15 are still within their deadlines. The overdue share is 5 ÷ 20 × 100: 25%.
The team reviews a sample of 20 closed responses. If 15 meet every agreed criterion, the quality rate for that sample is 15 ÷ 20 × 100: 75%.
The team also receives 20 completed surveys. If 16 respondents say they're satisfied, satisfaction among respondents is 16 ÷ 20 × 100: 80%. Report how many surveys were sent to help interpret that figure.
A fast response can leave a question unresolved. The Bogotá General Secretariat's model separates timeliness from quality and assesses criteria such as clarity and coherence.
A Minciencias metric sheet calculates satisfaction as satisfied people divided by survey respondents. Define the question and which answers count as satisfied before sending the survey.
Not every organization uses the same denominator. The Cundinamarca Institute of Infrastructure and Concessions uses a formula based on received cases. It notes that some cases are still within their deadlines at the cutoff date. Don't compare that rate directly with the on-time closure rate.
How do you establish a baseline before introducing AI?
A baseline records current performance using rules you can repeat after the change.
- Create one row per case. Record the ID, type, channel, and date received. Add the owner, status, deadline, and final response date. Include the reopening date when applicable.
- Add review data. Record the quality result and criteria used. Track surveys sent, responses received, and satisfaction results separately. Limit access to personal data.
- Review a sample of responses. Include different PQRS types and channels. Define how you'll select the sample. Have a person check whether each response addressed the request. Record incomplete or confusing responses too.
- Calculate and document. Keep the formulas, reporting period, and cutoff date. Note missing data and any changes to the criteria. This helps you explain a change before attributing it to a tool.
- Compare results after the change. Use the same definitions and review both timeliness and quality. If you use AI to classify cases or suggest responses, evaluate its output against human-reviewed examples.
Which tool should you use to track the metrics?
The choice depends on where cases arrive and how reliable your existing data is.
| Option | When it makes sense | What to check |
|---|---|---|
| Current tool | It already records dates, statuses, and owners consistently. | Exports, available fields, and permissions. |
| Automation platform | You need to consolidate data or send deadline alerts across tools. | Connections, errors, and workflow maintenance. |
| Custom development | Your systems or tracking rules require custom features. | Integrations, testing, and ongoing operation. |
Cost and timelines depend on record quality, required connections, and access controls. Calculation rules and maintenance also matter. To compare support tools, see our guide to help desks for PQRS. If pending cases are the problem, read about managing unanswered complaints.
