What is Resolution Rate?
Resolution rate is the percentage of customer support inquiries that get fully resolved out of the total received. A ticket counts as resolved when the customer's issue is actually fixed, not when it is closed, deflected, or abandoned. Closing a ticket and solving a problem are two different outcomes, and the metric only works when it tracks the second one.
In AI-driven support, resolution rate usually refers to the share of conversations an AI agent handles end to end without a human stepping in. It is related to, but stricter than, first contact resolution, which measures whether an issue was solved in a single interaction regardless of who handled it.
Vendors define resolution differently, so buyers should always ask how a published number was measured. Some count every conversation a bot answered; rigorous definitions require the customer's issue to be confirmed solved with no reopened ticket and no escalation.
Why Resolution Rate Matters
Resolution rate is the metric that connects support automation to money. If an AI agent resolves 90% of inquiries instead of 30%, cost per ticket, headcount plans, and queue lengths all change. Teams that benchmark vendors on resolution rate rather than vague automation claims make better purchasing decisions, which is why frameworks for measuring AI support ROI anchor on it.
It also reflects customer experience directly. A low resolution rate means customers bounce between a bot and human agents, repeating themselves at every step. That drags down CSAT and inflates contact volume, because unresolved customers always come back.
For executives, resolution rate is the single number that says whether AI support is working. Containment can be gamed by making escalation hard; honest resolution cannot.
How Resolution Rate Works
The formula is simple: resolved inquiries divided by total inquiries, multiplied by 100. The hard part is defining "resolved." Strong implementations require three conditions: the answer addressed the actual question, the customer did not escalate, and the ticket was not reopened within a set window, typically 24 to 72 hours.
Resolution rate is often confused with deflection rate, which counts conversations kept away from human agents whether or not the customer was helped. A bot that frustrates people into giving up scores high on deflection and low on honest resolution. Mature evaluation methods for containment and resolution quality treat the two as separate metrics for exactly this reason.
Benchmarks vary by industry and ticket complexity. Generic FAQ bots typically resolve 20 to 40% of inquiries, while platforms compared on resolution-rate performance score far higher by taking actions like refunds and account changes, not just answering questions.
How Fini Approaches Resolution Rate
Fini's autonomous AI agents deliver a 90% resolution rate across 3M+ monthly resolutions, measured on issues actually solved rather than conversations contained. Pricing follows the same logic: plans carry a monthly resolution allowance and charge per resolution beyond it, so you pay for outcomes instead of seats or message volume.
Fini backs the metric with a Zero Pay Guarantee: if it does not achieve 80% resolution in 90 days, you pay $0. Teams go live in 30 days with 99% accuracy across voice, chat, and email. To see resolution rate measured against your own ticket history, book a demo.
What does resolution rate mean in customer support?
Resolution rate is the percentage of support inquiries that get fully solved, whether by an AI agent or a human team. A conversation only counts when the customer's actual problem is fixed, with no escalation and no reopened ticket. It is the most direct measure of whether a support operation, automated or not, is doing its job.
How do you calculate resolution rate?
Divide the number of fully resolved inquiries by total inquiries received, then multiply by 100. The critical step is defining "resolved" strictly: the answer addressed the real question, the customer did not escalate, and the ticket stayed closed for a set window, usually 24 to 72 hours. Loose definitions inflate the number and hide failures.
What is a good resolution rate for AI customer support?
Generic FAQ chatbots typically resolve 20 to 40% of inquiries. Modern autonomous agents that take actions, like processing refunds or updating accounts, score much higher. Fini holds a 90% resolution rate across 3M+ monthly resolutions and backs it with a Zero Pay Guarantee: below 80% resolution in 90 days, you pay nothing.
What is the difference between resolution rate and deflection rate?
Deflection rate counts conversations kept away from human agents, regardless of outcome. Resolution rate counts problems actually solved. A bot that frustrates customers into abandoning their issue scores high on deflection and low on resolution. Buyers comparing vendors should always ask which metric a claim refers to, since the gap between the two can be enormous.
How does resolution rate differ from first contact resolution?
First contact resolution (FCR) measures whether an issue was solved in a single interaction, by anyone. Resolution rate measures whether issues were solved at all, across the full volume of inquiries. An AI agent can lift both: it resolves routine issues instantly, which raises FCR, while handling enough total volume to push overall resolution rate up.
Can AI agents really achieve a 90% resolution rate?
Yes, when they can take actions, not just retrieve answers. Resolving a refund request means issuing the refund, which requires secure API access to billing and CRM systems. Fini reaches 90% resolution with 99% accuracy by executing those workflows autonomously, with PII Shield redacting sensitive data in real time and SOC 2 Type II controls underneath.