Resolution rate is the number most business owners look for first when they open their Aiva Chat dashboard, and it's also the number most likely to be misread.
It varies from one business to the next, and the honest answer to "why is mine lower than theirs?" usually points to a decision made somewhere in setup that's working exactly as configured: a cautious handoff rule, a booking flow that isn't connected yet, or a price change nobody told the assistant about.
That matters because a number you misread leads to the wrong fix. An owner who treats a low resolution rate as an AI failure might loosen handoff rules that were protecting customers, when the real cause was a missing answer about a new service.
If your number looks low, skip ahead to the section on why it varies between businesses; most owners recognise their own setup in one of the four causes. The use cases further down are composites, not named customers.
What is the resolution rate dashboard?
It's the part of Aiva Chat's analytics that shows how many of your phone calls were handled without a person stepping in. On the Analytics page, the Results section includes a card called Resolved Without a Person, alongside:
- Appointments Booked
- New Callers Captured
- After-Hours Calls Handled
- Hours Saved and Money Saved (once you set hours and a staff cost)
- Handoff Rate
Together they answer a simple question: what did the assistant actually accomplish for the business over a given period?
Resolution here is about phone calls. Each call gets an outcome after it ends, and that outcome is what the card counts. Website chats are saved under Web Conversations, and text threads live in the SMS inbox, so neither feeds this number.
It's built for owners and managers who want an honest read on their phone line without becoming analysts. If you want the general definition and where it runs out, our companion piece on what resolution rate means covers the term itself.
How resolution is measured
Step 1: The call ends and is saved
Every call is stored with its recording, transcript, and summary. Nothing about resolution is decided while the call is still in progress.
Step 2: The transcript is reviewed and given an outcome
Shortly after the call, Aiva Chat reads the transcript and records what happened: resolved, needs follow-up, transferred, voicemail, no answer, or other. It also notes topics, the caller's intent, caller sentiment, and a quality score from 1 to 5.
Step 3: Outcomes are counted for your date range
The Results cards count those outcomes over the period you're viewing. Resolved calls feed Resolved Without a Person. Transferred calls, including transfers nobody answered where the AI then took a message, feed the Handoff Rate.
Step 4: Weekly findings and the Monday email
Each week, Aiva Chat lists what it found in your calls under What your calls are telling you, and on Monday morning emails a summary covering calls handled, calls resolved without a person, after-hours calls, bookings, hours saved, and up to three findings. The review calls and results guide walks through each screen.
What counts as resolved, and what doesn't
A concrete outcome, not just an ending
A call is resolved when the caller's need was handled: a question answered correctly, an appointment booked, moved, or cancelled, a status check completed. A call that simply ends doesn't qualify.
Transfers and voicemail don't count
A call handed to a person is, by definition, not resolved without a person. Neither is a call that went to voicemail or ended with a follow-up task someone now has to do.
Why we kept the bar strict
It's tempting to build a metric that rewards "the AI kept talking until the caller went away." We deliberately didn't. A number that counted every ended call would look great and mean nothing.
Some of the calls that never count as resolved are exactly the ones you decided a person should always take.
Why resolution rate varies between businesses
How much the assistant can actually do
Businesses with the highest resolution rates usually gave the assistant real capability, not just a document. If booking is set up end to end, "do you have anything Friday?" gets checked, offered, and confirmed. With only a short FAQ, more calls end in "let me have someone follow up," which is the right outcome but a lower number by design.
How cautious the handoff rules are
A business with cautious escalation rules, sending anything with a hint of complexity to a person, will show a lower rate than one that lets the assistant handle more. Neither is wrong; they're different choices about risk.
Information that lags behind the business
The most common fixable gap is a service or price that changed last week but never made it into the assistant's information. It either gives the old answer or correctly admits it doesn't know. Neither is a resolution.
Questions nobody planned for
A clinic set up for booking and hours may start getting insurance questions it was never given information about. Those calls correctly hand off, and stay outside the resolved count until the business decides to bring that topic in scope.
Benefits of tracking resolution rate
A signal while it's still fresh
Because outcomes are recorded shortly after each call, a drop shows up while it's still fresh enough to explain, like a new service added last week.
It points at fixable causes
Read next to topics, the number shows which category of call is dragging the average down. That's usually a specific missing answer, which takes minutes to fix.
It keeps handoff decisions honest
Seeing resolution and handoff rate side by side makes the trade-off visible. You decide deliberately how much to automate instead of guessing.
It quantifies time saved
With business hours and a staff cost set, Hours Saved and Money Saved translate resolved calls into terms an owner can weigh against the plan price.
Resolution rate use cases
The clinic with a sudden dip
Problem: Resolved calls drop over a week with no obvious reason. How the dashboard helps: Topics show a spike in calls about a newly added service. Opening a few shows the assistant saying it doesn't have details. Outcome: The owner adds the service information, and the next matching calls resolve.
The law firm that wants a low number
Problem: A firm's resolution rate is lower than a salon's they heard about. How the dashboard helps: Handoff rate shows most unresolved calls are transfers triggered by the firm's own rules for legal questions. Outcome: The firm keeps its rules. The number reflects a deliberate choice.
The salon with booking half-connected
Problem: Many calls end in "someone will call you back" about availability. How the dashboard helps: The outcome filter shows needs-follow-up calls clustered around booking. Outcome: Connecting the calendar properly turns those into booked appointments.
The repair shop getting off-script questions
Problem: Callers keep asking about warranty terms the assistant was never told. How the dashboard helps: "Ask your calls" answers a plain question, "what do people ask that we can't answer?", with links to the calls. Outcome: The shop adds a warranty section and watches that topic improve.
Resolution rate vs. other call metrics
| Metric | What it tells you | What it misses |
|---|---|---|
| Answer rate | Whether calls were picked up | Whether anything got done |
| Resolution rate | Calls handled without a person | Whether the caller was happy |
| Handoff rate | Share of calls sent to a person | Whether the handoff was the right call |
| Quality score (1 to 5) | How well the AI handled each call | Business outcomes like revenue |
| Caller sentiment | How the caller seemed by the end | Whether their need was met |
| Appointments booked | Bookings that came from calls | Calls that resolved without booking |
Answer rate is easy to make look good, since picking up isn't the hard part. Resolution rate is the more honest measure of work done, but it isn't a satisfaction score: a call can resolve cleanly and still leave a caller lukewarm.
That's why we keep these signals separate instead of folding them into one percentage. Collapsing them makes each harder to read. The U.S. government's customer experience program approaches it from another angle, describing service improvement as understanding customers' needs through feedback, research, and testing new approaches.
Read resolution with handoff rate and topics, and use quality scores and sentiment to spot calls worth listening to. Our guide to AI receptionist KPIs covers how these fit together.
How to use the dashboard well
Step 1: Set your hours and staff cost
In Business Settings, set open days and hours and your staff cost per hour, so after-hours and money-saved cards have what they need.
Step 2: Establish a baseline
Watch the first few weeks without changing much. Your own trend matters more than anyone else's number.
Step 3: Read it with topics
Look at which topics are most common and which ones appear among unresolved calls. The topic tagging behind this is automatic.
Step 4: Open the unresolved calls
Filter the call log by outcome or Needs review. Read the transcripts of a handful and note the pattern.
Step 5: Fix the answer, then re-check
Use Fix this answer to save a correction or add knowledge. Check the same topic the following week.
Common mistakes reading resolution rate
Comparing your number with another business's
Different setups, different handoff rules, different callers. A comparison across businesses tells you almost nothing. Compare your number with your own last month.
Loosening handoff rules to raise the number
If calls are handing off because your rules say they should, removing those rules raises the number and the risk. Change rules only when you're confident the assistant should handle those calls.
Reading the headline without topics
An overall rate that looks fine can hide one category resolving badly. The average never shows that on its own.
Treating it as satisfaction
A resolved call isn't automatically a happy caller. Check sentiment and listen to a few calls each week.
Counting chat and SMS in your head
The card covers phone calls. Review website chat conversations separately so you don't assume they're reflected.
Resolution rate best practices
- Track your own trend rather than chasing a target.
- Read resolution next to handoff rate so the trade-off is visible.
- Check topics weekly for categories that resolve poorly.
- Update information the day it changes: prices, services, hours.
- Use Fix this answer whenever a transcript shows a wrong reply.
- Keep handoff rules for sensitive topics, even if they lower the number.
- Read the Monday email and act on at least one finding each week.
Conclusion
The single most useful habit with this dashboard is to compare your number only with your own history. A business with strict handoff rules for legal or billing questions will always sit lower than a salon booking haircuts, and both can be running exactly as they should.
The decision point comes when the number drops and you can't say why. That's worth an hour: check topics for a new spike, filter the call log to unresolved calls, read five transcripts, and fix the missing answer. Most unexplained drops trace back to something that changed in the business and never reached the assistant.
The caveat is scope. The card covers phone calls only, and it doesn't tell you whether callers were happy. Read it alongside handoff rate and sentiment, and review website chats separately. We built it strict on purpose; a resolved call here means the caller actually got what they needed.
Ready to see what your calls are really telling you?
The fastest way to understand this dashboard is to watch it fill with your own calls. Set your business hours and staff cost first so every card has what it needs, then let a couple of weeks of calls build a baseline before changing anything.
After that, open the topics that resolve poorly and fix one answer a week. Analytics comes with every plan, Starter included at $49 a month.
Start your 14-day free trial with no card required, or book a demo to read a sample dashboard with us.
Posts from the people building Aiva Chat in Rajkot — how the product works, what we have learned building it, and guides for getting the most out of an AI receptionist.