SupportIQ

What SupportIQ does

SupportIQ grades every support case, for every agent and bot on it, against your quality standards. You use it to find coaching opportunities, coach agents, and check grade accuracy.

• Grades 100% of cases. New cases are graded automatically. History is backfilled at setup.

• Two high level metrics: Quality Performance Rate (Percentage of cases where the agent met our standards) and inferred customer satisfaction (DSAT, the percentage of cases where the customers was dissatisfied or very dissatisfied).

• Coaching: focus areas per agent, backed by case evidence, with notes, goals and history.

• Human review: correct any grade, or run structured calibration samples.

• SOP-driven: grades follow your SOPs, and calibration finds SOP gaps and recommends changes or additions.

Roles and access

Roles stack: one person can hold several. Your admin assigns them.

Leadership (directors, VPs) usually holds Support Manager access scoped to the whole org.

Getting around

Log in at app.unwrap.ai with your work email, through your company's SSO if it uses one.

Sidebar. Everything lives under Support IQ:

Set your homepage

1. Open Performance.

2. Click into your team - if there's no breakdown, let us know how your teams are grouped so we can select the proper breakdown. You can use multiple breakdowns (e.g., Team, Region, Queue, etc).

3. Optional: bookmark you team as your homepage, or keep the org-wide view as home. SupportIQ will open to your bookmarked team going forward if selected.

4. Bookmark the URL too. It keeps your filters and date range.

Key terms

Why QPR and DSAT are separate: a customer can be unhappy because of a product or policy issue the agent couldn't fix. QPR measures the agent; DSAT measures the customer, but customer experience is heavily influenced by Agent behavior. Agents high on both DSAT and Needs Improvement are the highest priority.

You can also filter for cases where agents turned customers from dissatisfied to satisfied over the course of the conversation using the "Support Recovery Flag" filter.

How grading works

1. A case comes in from your support platform and is graded automatically.

2. Each participant is graded separately. Every human agent and bot on the case gets their own grade.

3. The base rubric checks universal behaviors, like answering every part of the customer's question, staying consistent, acting respectfully, not making customers repeat themselves.

4. Your SOPs add company-specific standards, such as your quality scorecard, policies and troubleshooting steps. SOPs can apply to all cases or only to certain case types, channels, or to bots or humans only.

5. The result is pass/fail: Satisfactory or Needs Improvement. There are no partial scores. Missing any rule means Needs Improvement. Ambiguous cases default to Satisfactory.

6. The grade summary explains the grade and cites the case. When an SOP drove the grade, the SOP is cited.

7. Corrections happen on the case or in manual QA, and allow a manager or QA reviewer to examine the grade and correct it with rationale.

Key points

• Grades appear immediately. They are not held for review.

• The AI only knows what the rubric, SOPs and case text tell it. Incorrect grades can be improved by closing policy context gaps.

• Corrections are logged and used to recommend SOP improvements or additions.

• To see the rules, click View grading rubric. To see the SOP behind a specific grade, open the case in calibration and click View SOP used in grading this case.

For agents

You see only your own page. Use it to check your grades and follow your coaching.

See your performance

• Your QPR and DSAT tiles, with change versus the prior period.

• Your QPR chart against the team average (dashed line), plus your goal if your manager set one.

• Your focus areas: the gaps that show up most in your Needs Improvement cases, with example cases.

Review your cases

1. Open the All cases tab.

2. Filter to Needs Improvement to see what to work on.

3. Click a case to read the conversation and the grade summary.

Coaching notes

• Notes your manager shares appear in Coaching history.

• Read each note and mark Acknowledgment, if available, once reviewed. Your manager sees when you acknowledged it.

Request a regrade

Think a grade is wrong? Request a regrade on the case and explain why. Your manager approves or declines it.

For managers: assess your team

Read the team page

1. Open Performance (or your homepage). Set the date range and filters.

2. Check the three tiles: Cases Graded, DSAT Rate, Quality Performance Rate.

3. Scan the QPR chart for dips.

4. Read Ways Support Can Improve for the top team-wide gaps. Use them in team meetings and training.

Find agents who need attention

1. Go to Team Members. Sort by QPR, DSAT Rate or 30d change.

2. Flag agents well below the team average, trending down, or with no recent Last Coached date.

3. Use Add metrics to add columns such as time to resolution or reopen count.

4. Use the Handled Cases tab to see every team case in one list.

Open an agent

1. Click the agent's row. The Coaching tab opens.

2. Compare their QPR line to the team average.

3. Switch to All cases and filter by inferred CSAT and grade to pull the cases that matter.

4. Click a case to read the conversation and grade summary. The source icon opens the case in your support platform.

For managers: coach an agent

Find the performance gap, track it, coach and log it, check whether it moved. It all happens on the agent's Coaching tab.

1. Find the gap

• Suggested focus areas are ranked by how often each appears in the agent's Needs Improvement cases. The percentage is of Needs Improvement cases, not all cases.

• Click a focus area to see the cited cases. Read two or three before coaching.

2. Track it

• Pin up to 3 areas to Tracked focus areas, or click + Add your own to write one. SupportIQ finds matching cases for custom areas.

• Always pin what you coach on. Unpinned suggestions refresh and may change.

3. Set a goal

• Click Set QPR goal on the agent's chart. The goal shows as a line.

4. Log the coaching

1. Write a coaching note: what you discussed and what the agent committed to.

2. Set the coaching date. It can be backdated.

3. Choose whether to share it with the agent. Save.

4. The note appears in Coaching history and Last Coached updates.

5. The agent acknowledges shared notes. You see the timestamp.

Use notes to mark when a coaching area starts and ends. Adding or removing a tracked area is also timestamped.

5. Prep 1:1s with a coaching report

1. Click Create shareable report. A coaching report opens with the agent's focus areas, cases and coaching history.

2. Edit it or refine it in the report chat. Save and share.

3. Set it to recur (for example, weekly before your 1:1). It arrives by email with week-over-week changes.

4. Save a version as a coaching template to reuse it.

6. Check whether it worked

• A week later, reopen the tracked focus area. Is the share of cases with that gap going down?

• Check QPR and 30d change. Rising QPR is the goal.

• Not there yet? Log a follow-up note and keep it tracked.

Reviewing grades

There are two ways to review grades. Both use the AI grade as the starting point; you confirm or override it.

Manual review

1. Open an agent's All cases tab. Filter to Needs Improvement or Dissatisfied.

2. Each case card shows Inferred CSAT and Agent performances (one grade per participant).

3. Click into the case. Each agent's grade sits beside their messages, with the Grade summary below.

4. Agree? Move on. Disagree? Correct it (see next section).

Grading Calibration: create a sample

1. Go to Grading Calibration. Click + New sample.

2. Name it, set the size, and add filters (date range, grade, team, exclude bots).

3. Add reviewers. Add several to the same sample to compare reviewers on the same cases.

4. Save. The sample shows its filters, size, progress, correction rate and reviewers.

Grading Calibration: review a sample

1. Open the sample. Review queue (left) lists cases left. Recently reviewed lists your history.

2. Read the conversation (middle). Click View SOP used in grading this case if needed.

3. In Agents in this case (right), each participant shows the AI grade and explanation. The person you are grading is highlighted in blue.

4. Click Agree or Disagree for each participant.

5. On Disagree, pick the correct grade and enter a reason (required).

6. Click Submit and next. Use Agree with remaining to accept the rest. Use Skip case to pass.

Finish a case before opening another. Switching mid-case discards your verdicts on it.

Shortcuts: A = Agree, D = Disagree, 1 = Satisfactory, 2 = Needs Improvement, S = Skip, up/down arrows = switch agent. An undo shortcut is shown at the bottom of the panel.

Results

• Correction rate: share of grades reviewers disagreed with. Aim for low single digits.

• Stats: progress and results for a sample.

• Reviewers table: samples finished and grades reviewed per person.

• Export CSV (one sample) or Export grade changes (all samples): case link, original grade, new grade, reason, reviewer, timestamp.

The AI grade is visible during review by design. It often catches details reviewers miss. Blind review can be enabled on request.

Correcting grades

Both views change the same grade and write to the same audit log.

Write useful reasons. Name the rule the AI missed.

• Good: "Agents may [action] in this situation per [SOP name]."

• Not useful: "This was fine."

After a correction

1. The new grade replaces the AI grade and updates QPR.

2. The change is logged: original grade, new grade, reason, who, when.

3. Corrections drive suggested SOP updates. Once approved, cases can be re-graded so the same error stops.

Regrade requests from agents appear for the agent's manager to approve or decline.

For leadership

Use the org view to see where quality is slipping, whether coaching is happening, and whether grades can be trusted.

• Org view: open Performance at the org level. Break down by team, manager, queue or tier to compare groups.

• Top gaps: Ways Support Can Improve shows the most common org-wide issues. Use it to set training priorities.

• Coaching activity: sort Team Members by Last Coached to see who is overdue.

• Recurring reports: subscribe to reports by email (for example, weekly QPR and DSAT by team) so you don't have to log in to stay current.

• Coaching roll-ups: Unwrap can set up reports such as coaching compliance (did each agent get their required coachings) and QPR change after coaching, by manager.

• Grade accuracy: check correction rates in Grading Calibration. Use Export grade changes for a full audit log.