QA managers and contact center team leads often focus their evaluation form redesign efforts on scoring methodology while underinvesting in form structure. A well-designed evaluation form does more than produce scores. It generates the transcript evidence needed for coaching, the consistency needed for calibration, and the data structure needed for automation. This guide covers how to build a customizable QA evaluation form and how to automate it at scale.
What are the 4 pillars of QA?
The four pillars of QA in contact centers are quality planning, quality control, quality assurance, and quality improvement. Quality planning defines the criteria and standards. Quality control applies those standards to evaluated calls. Quality assurance validates that the evaluation process itself is consistent. Quality improvement uses evaluation data to drive training and process changes. A well-designed QA form supports all four: it defines what is measured, ensures consistent measurement, and produces the data that feeds coaching and improvement workflows.
What are the quality assurance metrics for call centers?
Core QA metrics include overall call score, criterion-level scores by category (compliance, quality, soft skills), first call resolution, script adherence rate, escalation rate, and agent score trend over time. The most actionable metrics are criterion-level scores because they identify the specific skill gap rather than an aggregate number. An overall score of 72% tells you there is a problem. A compliance disclosure score of 45% tells you what the problem is and where to route the coaching.
Sample QA Evaluation Form
Before customizing a form, it helps to see the structure you are working with. Below is a baseline template that covers the core categories most contact center QA programs require.
| Criterion | Weight | Score (0-100) | Evidence (Transcript Quote) |
|---|---|---|---|
| Greeting and ID verification | 10% | ||
| Discovery and needs assessment | 20% | ||
| Objection handling | 20% | ||
| Compliance disclosure | 25% | ||
| Close or next steps | 15% | ||
| Soft skills (empathy, pace, tone) | 10% | ||
| Overall weighted score | 100% | ||
| Coaching notes |
The evidence column is the differentiator between a form that produces scores and one that produces coaching. Requiring evaluators to quote the transcript prevents the grade inflation that occurs when scorers evaluate from memory rather than evidence.
Step 1: Define your scoring categories
Start by mapping your categories to the business outcomes you actually care about. Most contact center QA forms use three parent categories: compliance, quality, and soft skills.
Compliance covers items where the agent must say specific things to meet regulatory or policy requirements. Verbatim or near-verbatim requirements belong here. Non-compliance should carry a higher weight because the business risk is higher.
Quality covers the substance of the interaction: whether the agent asked the right questions, handled objections accurately, and guided the customer to a clear outcome.
Soft skills covers delivery: tone, empathy, pacing, and professionalism. These often carry lower weights than compliance and quality but have an outsized effect on customer perception.
Insight7 supports a weighted criteria system with main criteria, sub-criteria, and context descriptions. Weights are configurable and must sum to 100 percent. Categories are editable at any time, so the form can evolve as your standards change without requiring a platform reconfiguration.
Step 2: Choose verbatim versus intent-based scoring per criterion
Not all criteria should be scored the same way. Compliance disclosures require verbatim or exact-match evaluation. Discovery questions should be scored on intent, because reps who ask "what brought you in today?" and "what are you trying to solve?" are doing the same thing even if the phrasing differs.
Making this distinction criterion by criterion prevents two common problems: compliance items being scored leniently based on intent, and conversational items being scored harshly because the exact script phrase was not used. Insight7 provides a toggle per criterion for script-based versus intent-based evaluation, which means the form can enforce precision where it matters and allow natural language where it does not.
Avoid this common mistake: Applying intent-based scoring to compliance disclosures because "the rep covered it" even when the required language was absent. Regulators evaluate verbatim compliance, and your form should too.
Step 3: Set criterion weights
Weights should reflect the actual business priority of each criterion, not an equal distribution for convenience. A form where every criterion has a 10% weight treats a missed compliance disclosure the same as a slightly abrupt close, which does not reflect the difference in business risk.
A useful starting point: compliance criteria combined should represent 30 to 40 percent of the total score, quality criteria 40 to 50 percent, and soft skills 10 to 20 percent. These ranges will vary by industry and call type. An ICMI resource on contact center quality management notes that quality monitoring effectiveness depends heavily on criteria design and weighting, not just evaluation frequency.
Recalibrate weights quarterly by reviewing which criterion scores correlate most strongly with CSAT or FCR data. If compliance score has no correlation with outcomes and empathy score has a strong one, the weights need rebalancing.
Step 4: Add evidence fields
A score without evidence is an opinion. Every criterion on the form should have a field for the evaluator to paste the specific transcript quote or timestamp that justifies the score.
This requirement does three things. It forces the evaluator to review the actual transcript rather than score from impressions. It creates an audit trail that agents can review during calibration sessions. It generates the evidence base for coaching conversations, so the manager arrives at the session with specific quotes rather than general observations.
Insight7 surfaces the exact transcript quote and call location for every criterion score automatically. The evidence field is pre-populated by the platform, removing the manual step of identifying and pasting quotes per criterion.
Step 5: Include a coaching assignment field
A QA form that produces scores without routing them to action is a data-collection exercise, not a performance improvement system. Adding a coaching assignment field at the bottom of the form closes the loop between evaluation and development.
The field should record two things: which criterion drove the coaching assignment, and which practice scenario or training content was assigned as a result. This creates a traceable record that connects a specific score gap on a specific call to a specific training action, which is necessary for measuring whether coaching is working.
Step 6: Automate scoring with AI to remove manual review dependency
Manual QA teams typically evaluate 3 to 10 percent of calls, which means most agent behavior goes unobserved and most coaching opportunities are missed. Automating scoring with AI converts the form from a spot-check instrument into a full-coverage system.
Insight7 applies your configured criteria and weights to 100% of calls, generating per-agent scorecards, criterion-level trends, and evidence-backed evaluations without manual listening. The platform reaches 90 percent or greater scoring accuracy after four to six weeks of calibration, at which point automated scores align with supervisor judgment closely enough to drive coaching decisions. According to SQM Group research on automated versus manual QA, manual review limits coverage to roughly 1 to 2 percent of interactions, making pattern-level analysis unreliable at the agent level.
The transition from an Excel-based form to an automated platform does not require rebuilding your criteria from scratch. Import your existing categories and weights, calibrate the scoring context, and the platform applies the form to every call your recording infrastructure captures.
FAQ
How many criteria should a call center QA form have?
Most effective QA forms use between five and eight scored criteria. Fewer than five produces scores that are too coarse to drive targeted coaching. More than ten creates evaluation fatigue, grade inflation, and inconsistent scoring across evaluators. Six to eight criteria covering compliance, quality, and soft skills covers the performance dimensions that matter most without overwhelming the evaluation process.
How often should QA forms be updated?
Review the form quarterly and update it when business priorities shift, when calibration data shows persistent evaluator disagreement on specific criteria, or when new compliance requirements are introduced. Updating weights is faster than adding new criteria and often produces better results. If a criterion consistently scores near 90% across all agents, it is either not discriminating between performance levels or not measuring the right thing.
What is the difference between a QA form and a compliance checklist?
A compliance checklist is binary: the agent either completed the required action or did not. A QA evaluation form is multi-dimensional: it scores both compliance and quality criteria on a graduated scale and produces a composite score that reflects the full range of agent behavior. Compliance checklists are useful for regulatory purposes but insufficient for coaching because they do not capture the quality and soft-skill dimensions that drive customer experience outcomes.
