Table of Contents
Key Insights
What Is SupportIQ?
SupportIQ by Unwrap is AI quality assurance (QA) and coaching for customer support teams. It grades every conversation, from human agents and AI agents, on how well each case followed your processes and served the customer. It shows managers what to coach and measures whether the coaching is working.
It also traces recurring problems to their root cause, so your team can prevent them instead of handling them again and again. Every repeat ticket costs agent time and customer patience.
In short: grade every case, coach every agent, and fix what's behind the frustration.
Why Traditional Support QA Falls Short
Most support QA programs grade a small sample of conversations against a fixed scorecard. Teams typically review just 2–5% of conversations by hand, according to Zendesk. Managers deliver generic coaching on whatever the sample flagged, then hope it sticks.
That approach leaves 4 gaps. Here's how manual sample QA compares with SupportIQ on each one:
Here's how that plays out. A QA team reviews a handful of tickets per agent each month, and managers coach on what got flagged. Meanwhile, a product bug is quietly driving a big share of contacts. Agents keep handling those tickets one at a time, and QA only looks at how each ticket was handled, so nobody connects them back to the bug.
How SupportIQ Grades Every Case the Way Your Best QA Lead Would
SupportIQ reads every conversation in full and grades it the way an experienced QA lead would: on whether the agent did the right thing for the customer, not just whether they followed a script. It runs on the same CX-trained AI that powers the rest of Unwrap. That natural language processing reads each conversation, sentence by sentence, to understand what the customer needed, what the agent did and how it ended.
It understands what your SOPs are for, not just what they say
Every SOP exists to get a certain result for the customer. SupportIQ grades whether the agent got that result, not just whether they followed each step in order.
Say your SOP is "offer a replacement before a refund." A customer has already had 2 replacements fail and just wants their money back, so the agent goes straight to the refund. A scorecard marks that as a violation, creating a headache for the agent and their manager. That's the opposite of what AI QA is meant to accomplish. SupportIQ sees that the agent met the goal of the SOP, to keep a good customer whole, and grades it that way.
It works in the other direction too. An agent who hits every scripted step (greeting, verification, empathy statement, closing) but never solves the problem passes a scorecard. SupportIQ flags it for not addressing the customer's true issue.
It grades against your own SOPs, from day one
SupportIQ comes with a universal grading rubric out of the box, so you see graded cases on day one.
Then you bring in your own standards. Add the support protocol and SOPs your team already uses, and SupportIQ grades every case against them:
- Your support protocol: the values, procedures and workflows that apply to every case.
- Situation-specific SOPs: the expected processes and edge cases for particular case types or customers, like refunds, chargebacks or VIP accounts.
Every grade shows its work
Each grade comes with the reasoning behind it, the parts of the transcript that support it and the SOPs it was graded against. QA leads can validate any score quickly, and agents can see exactly why a case was marked down.
It grades AI agents and human agents separately
This holds even inside the same conversation. If your chatbot sends a customer in circles before handing off, and the human agent fixes it quickly, the agent gets credit for the recovery while the bot gets flagged. A blended score hides the performance behind a single number.
It reads how every customer felt, not just the ones who answer a survey
Survey CSAT only hears from customers who respond, and customers who leave happy often don't bother. SupportIQ infers satisfaction from every support conversation instead, reading how the customer felt at each point and how they felt walking away.
It leads with DSAT rate, the share of customers who left dissatisfied, because frustration is the clearest signal of risk. And when an agent turns a frustrated customer into a satisfied one, SupportIQ flags the recovery, so great saves get recognized.
It helps you improve your SOPs, not just enforce them
Because SupportIQ connects what agents did to how each case turned out, it also shows when the process is the problem, not the agent. It does this in 4 ways:
- It tests SOPs against results. If agents who follow the SOP have worse outcomes than those who color outside the lines, the SOP needs to change, not the agents.
- It shows which steps frustrate customers. Because SupportIQ infers satisfaction on every case, you can see which steps in your process leave customers unhappy.
- It flags gaps when new problems appear. When issues show up that your current SOPs don't cover, SupportIQ recommends how those SOPs should change, or where new ones are needed.
- It learns from disagreements. When a new promo lets agents offer discounts the old SOP didn't allow, agents dispute those grades, SupportIQ recommends updating the protocol, and the QA lead signs off.
Every recommended change waits for your approval. Your SOPs only change when you say so.
How Do You Know If Agent Coaching Is Working?
With SupportIQ, you track progress on the specific skill you coached, not just an agent's overall score. That answers the question every support leader asks: is my coaching actually working?
Coaching that's ready for the 1:1
Managers don't have to hunt for the coachable moment. Before a 1:1, a manager opens an agent's page and sees their top improvement areas (say, not setting expectations on refund timing), with the transcripts to back them up. Coaching plans are drafted automatically, saving managers hours per week.
Progress on each improvement area
Overall scores can hide key focus areas. Take a manager who coaches an agent on refund-timing expectations. A month later, that skill has improved. But unless each case is read and graded on that specific improvement area, the manager has only the agent's overall score to go on, which may not show a change at all.
SupportIQ identifies an agent's most impactful improvement areas, then tracks progress along that personal growth path rather than against generic goals.
Coaching as a two-way conversation
Agents get their own view of their own data, in real time. They can see the goal their manager set, review example cases on their own time and track their progress between 1:1s.
Beyond QA: Coach the Agent, or Fix the Cause
Grading tells you how an agent handled a case. Customer intelligence tells you why the case existed, and that second part is where Unwrap comes in. Unwrap reads customer feedback from support tickets, surveys, app reviews, social posts and call transcripts, and automatically groups it into themes, with no manual tagging.
SupportIQ is built on the same platform, so support conversations land in the same themes as everything else your customers tell you. That's what lets it look past the individual case to what's causing it.
Say your agents are struggling with billing issues. That tells you where to coach. But maybe those issues are hard to resolve because a product problem is creating them. SupportIQ traces the billing issue back to its cause and gives the product team the evidence to fix it.
Improvement themes across the whole org
SupportIQ ranks the biggest improvement themes across your org, in plain language, by impact. If the top theme is agents repeatedly resending a verification link without figuring out why it fails, that's partly coaching and partly a broken verification flow. You see both.
Tell an agent problem from a business problem
SupportIQ scores how the agent handled a case separately from how the customer felt about the outcome. When the two don't match, that's a signal. If agents handle late-delivery cases by the book but those customers are still unhappy, more coaching won't help. The problem is the carrier, and that belongs with your logistics team.
Prepare agents for emerging issues
Unwrap's real-time alerts flag new issues as soon as they start spiking in support. Say a pricing change starts confusing customers about their bill. The team can send agents guidance and a macro right away, so they handle those cases with confidence instead of frustrating customers.
The result: root causes go to the teams that own them, and fewer customers need to contact support at all.
Who Is SupportIQ For?
SupportIQ is built for large support operations, in B2C or high-volume B2B, including teams that work with BPOs.
- QA leads get full coverage, and grades they trust.
- Support managers get coaching that's ready for the 1:1, without extensive prep work.
- CX and support leaders get fewer tickets, higher CSAT and NPS, lower churn, and data to bring to product roadmaps.
Frequently Asked Questions
What is SupportIQ?
SupportIQ by Unwrap is AI quality assurance and coaching for customer support teams. It grades every conversation, from human and AI agents, against your own SOPs, tracks how coaching improves each skill, and traces recurring problems to their root cause so teams can prevent them.
How is AI QA different from manual QA?
Manual QA reviews a small sample of conversations against a fixed scorecard, usually a few per agent each month. AI QA grades every conversation automatically. Unwrap’s SupportIQ also understands what each SOP is meant to achieve, so it credits good judgment and flags agents who follow the script without solving the problem.
How accurate is AI QA, and can I trust the scores?
Every SupportIQ grade includes the reasoning behind it, the transcript evidence and the SOP it was graded against, so QA leads can check any score. Managers can correct grades and agents can request regrades, and those corrections improve future grading.
Does AI QA replace my QA team?
No. Unwrap’s SupportIQ takes sampling and manual scoring off your QA team's plate, so they can spend their time on the work that needs people: refining standards, reviewing edge cases and coaching agents.
Can SupportIQ grade AI agents and chatbots?
Yes. Unwrap’s SupportIQ grades AI agents and human agents separately, even within the same conversation. A bot's dead end doesn't drag down the agent who fixed it, and you can see where your AI agent needs work.
Can SupportIQ help improve my SOPs?
Yes. Unwrap’s SupportIQ compares SOP compliance with case outcomes, shows which process steps frustrate customers and flags issues your current SOPs don't cover. It recommends changes, and every change waits for your approval.
How long does SupportIQ take to set up?
Unwrap’s SupportIQ works on day one with a built-in universal rubric, and the free trial backfills your historical conversations, so you see results right away. Add your own support protocol and SOPs whenever you're ready. If you already send support data to Unwrap, there's nothing new to integrate. If not, hundreds of out-of-the-box integrations, including Zendesk, Intercom, Salesforce, Gladly, RingCentral, and Gong, get you running in hours.
