FCR has been the standard yardstick for support quality for decades, and the numbers behind it are stark. SQM Group, which has benchmarked over 500 North American contact centers for 25 years, puts the aggregated industry average at 71%. Only 5% of teams hit the "world-class" threshold of 80% or higher.
That gap is worth caring about because it compounds. Every 1% improvement in FCR tracks with a 1.4-point lift in NPS, per the same research.
Brainfish customer Vrio automated 80 to 85% of its developer support questions and cut daily support volume from 30 to 50 messages down to just the complex edge cases. Routine questions now route straight to grounded, always-current documentation instead of pulling an engineer into every thread.
How is first contact resolution calculated?
The formula is simple: issues resolved on the first contact, divided by total issues handled, times 100. A team that closes 710 out of 1,000 tickets without a second contact is running at 71% FCR, right at the industry average.
The hard part isn't the math. It's agreeing on what counts as "resolved" and what counts as "one contact," since a chat thread spanning three days can still count as a single contact if the customer never had to re-explain the problem.
What's a good FCR rate?
Anything from 70% to 79% is a healthy, "good" FCR rate. Below 70%, a team is trailing the average it's being compared against, while above 80% puts it in rare company, roughly 5% of the contact centers SQM benchmarks.
FCR also varies by issue type. Simple, low-complexity requests resolve on the first try far more often than technical or emotionally charged ones.
Source: SQM Group benchmark data by issue type.
If your queue skews technical, a "low" 63% FCR rate might actually be strong performance for that category, not a red flag.
Why does FCR matter more than other support metrics?
FCR ties customer experience directly to operating cost. A customer whose issue gets solved the first time doesn't call back, open a second ticket, or get more frustrated with each additional touch, and a 1% FCR improvement tracks with roughly a 1% reduction in operating cost, per SQM's research.
How is FCR different from resolution rate and deflection rate?
These terms get used loosely, and the differences matter when you're comparing numbers across tools or teams.
- First contact resolution (FCR): Solved in one interaction, no follow-up needed. The strictest of the three.
- AI resolution rate: The share of AI-handled conversations closed without a human, whether or not it took several messages in that one session.
- Deflection rate: The share of total volume that never reaches a human agent, which can include a customer who simply gave up.
FCR is the only one that requires the issue to actually close on the first attempt. The other two can look good while a customer is still stuck.
What causes a low FCR rate?
A handful of patterns show up again and again in support teams' own root-cause reviews:
- Agents can't find the answer. Knowledge is scattered across docs, tickets, and a few senior teammates' heads. It's one of the patterns SQM's own research flags most often.
- Policies block resolution. An agent knows the fix but has to escalate for approval.
- The request wasn't actually understood. The agent solved the wrong problem the first time.
- Handoffs lose context. A second agent restarts the conversation instead of picking it up.
Most trace back to one root cause: the agent, human or AI, didn't have the right information when the customer asked. The last point is worth dwelling on, because a bad handoff doesn't just feel worse to the customer, it usually creates the second contact directly. When a conversation drops into a fresh ticket queue with no context, that new ticket is, by most FCR definitions, a second contact on the same issue.
"AI handles the easy half of the question, the customer escalates, and the second half of the conversation becomes adversarial because the customer is now re-explaining context the AI already had. CSAT on escalated conversations sits 15 to 25 points below CSAT on AI-resolved or human-only conversations. The escalation itself is what damages the experience, not the AI's inability to resolve." Daniel Kimber, CEO & Co-founder, Brainfish, in Introducing Brainfish Live Agent Handoff for Zendesk
How do you improve first contact resolution?
Start with the knowledge an agent is working from, not the agent's training. If the answer isn't accurate, current, and easy to find, no amount of coaching fixes a wrong response.
Three changes actually move FCR:
- Consolidate scattered knowledge into one layer that continuously ingests docs, tickets, and Slack threads, instead of leaving each source to drift out of sync with the others.
- Give agents, human and AI, permission to resolve, not just diagnose, so a correct answer doesn't stall on an approval step.
- Track repeat contacts by reason, not just the blended FCR number, so you can see which topics are actually driving the second contact.
AI support runs into the same wall human agents do. An AI agent working from scattered, outdated sources fails the same way a poorly-trained new hire does, which is a large part of why MIT's 2025 State of AI in Business report found that 95% of enterprise generative-AI pilots never deliver measurable ROI.
Brainfish is built to close that gap: it grounds a support team's AI agent in real, continuously updated product knowledge and runs it alongside the helpdesk already in place, so the agent and the customer land on the same accurate answer the first time. For tactics specific to complex B2B software support, see our guide on improving first contact resolution in enterprise SaaS support.
Ready to see what first contact resolution looks like when the knowledge behind it is actually current? See how Vrio automated 80 to 85% of developer support, check pricing, or book a demo.