Contents
Every insurance broker knows the feeling.
Monday morning. You arrive at the office. The voicemail light is blinking. Fourteen messages from the weekend. Three are new business enquiries, but one already says they went with someone else after calling you Saturday and getting no answer.
Friday 5:30pm. The phones have been switched to voicemail. Your mobile rings. A customer's had an accident and needs to report a claim. You take the call because you are still there, but you know if you had left five minutes earlier, they would have reached voicemail too.
Bank holiday. The office is closed. Your competitors who have call centres are open. You are not. Every enquiry that weekend goes to someone else.
This is the reality of traditional insurance broking. Limited hours. Missed calls. Lost opportunities. The business works when you are available and stops when you are not.
One broker decided to change this.
The Problem
This particular broker, a specialist motor operation with about fifteen staff, was processing solid volume during business hours. Quote requests, policy changes, renewals, claims. The team was busy and effective.
But analysis revealed a problem hiding in plain sight.
Voicemail volume: Average of twelve voicemails per day, higher on Mondays (weekend accumulation) and after holidays.
Callback success: Of voicemails requesting callback, only 60% were successfully reached. The rest went to voicemail themselves, starting phone tag that sometimes never resolved.
New business in voicemail: Approximately 30% of voicemails were new business enquiries. At their average premium and close rate, each missed new business call represented roughly £180 in potential premium.
Weekend pattern: Call attempts (measured by voicemail + web form submissions timestamped after hours) averaged 25-30 per weekend. That is over 100 per month, during hours when the business was closed.
The hidden cost: Conservative estimate: fifteen lost new business opportunities per month at £180 each. That is £2,700 monthly, £32,400 annually, just in identifiable new business loss. Customer service calls that went to voicemail and never called back, renewals that went elsewhere after failed contact, claims reported late because customers could not reach anyone: all additional and harder to quantify.
The broker was running a successful business during business hours and bleeding opportunity every evening, weekend, and holiday.
The Decision
The obvious solutions did not fit.
Night shifts: For a fifteen-person operation, staffing evenings and weekends would require at least two additional full-time equivalents, more accounting for coverage requirements. At £25,000-30,000 each plus on-costs, that is £60,000-75,000 annually for after-hours coverage. The maths did not work.
Outsourced call centre: Third-party services could answer calls, but they could not access systems, could not quote, could not process changes. They would take messages, which was little better than voicemail. Plus they lacked specialist motor knowledge. Customer experience would suffer.
Extended hours: Keeping the office open until 8pm and opening Saturdays might capture some additional volume, but it would not capture late evenings, Sundays, or the middle of the night. And it would require premium wages for antisocial hours.
AI voice agents: A system that could answer calls, understand customer needs, access systems to provide information, capture details accurately, and escalate to humans when needed. Operating 24/7/365 at a fraction of the cost of human staffing.
The broker chose AI.
The Implementation
Deployment happened in phases over two months.
Phase 1: After-Hours FNOL (Week 1-3)
First priority: capture claims calls after hours. Accidents do not respect business hours. A customer unable to report a claim might call a competitor or claims management company instead.
The AI voice agent was configured to:
- Answer calls immediately
- Identify FNOL calls through natural conversation
- Capture incident details (date, location, circumstances, third party info)
- Verify policy details where possible
- Create claim records in the broker's system
- Send confirmation to the customer and alert the claims team
First weekend: four FNOL calls handled entirely by AI. All four had accurate information captured. All four customers expressed relief at reaching someone.
Phase 2: Status and Information Queries (Week 3-5)
Next: the "where's my claim?" and "what's my policy number?" calls that require no human judgement.
AI was connected to policy and claims systems with read access. It could now:
- Verify caller identity
- Retrieve policy information
- Provide claim status updates
- Confirm document receipt
- Explain next steps in processes
These queries previously generated voicemails that required callback and system lookup. Now they resolved in real-time, often in under two minutes.
Phase 3: Quote Capture (Week 5-7)
New business enquiries were the biggest opportunity. AI was configured to:
- Identify quote requests
- Gather risk information (vehicle, drivers, history)
- Capture contact details accurately
- Explain the quote process
- Set expectations for callback timing
The AI did not generate quotes. Pricing required human underwriting judgement. But it captured everything needed for the team to quote efficiently first thing next business day.
Phase 4: Full Deployment (Week 7-8)
Final phase: AI answering all calls when the office was closed, plus overflow during busy periods.
Integration with the dialer enabled seamless handover: if the office was open and handlers available, calls routed to humans. If the office was closed or all handlers busy, calls routed to AI.
The Results
Twelve months of operation produced clear metrics.
Call Capture
Before: ~60 calls per week reaching voicemail After: Zero calls reaching voicemail
Every call answered, every time. Evenings, weekends, holidays, 2am: all answered by AI.
After-Hours Resolution
AI-resolved queries: 68% of after-hours calls fully resolved by AI without human follow-up Human callback required: 32% required human follow-up, but with complete information captured
The 68% represented status checks, information queries, and confirmations that previously generated voicemails and callback attempts.
New Business Capture
Before: ~15 new business enquiries per month during closed hours, many lost After: 100% captured, with complete information for morning follow-up
New business callback success rate improved from 60% (when calling back voicemails with often incomplete numbers) to 95% (when calling back AI-captured leads with verified contact details).
Estimated additional new business: £35,000-40,000 annually in premium that would otherwise have been lost
Claims Handling
FNOL after hours: Average 12 per month captured by AI Information quality: Complete in 94% of cases (vs. ~60% for voicemails) Customer satisfaction: Higher than historical baseline for FNOL
Claims reported immediately rather than waiting for callback. Information captured while fresh. Customers less stressed.
Customer Feedback
Survey of customers who interacted with AI after hours:
- 87% satisfied or very satisfied with the experience
- 92% preferred AI answer to voicemail
- 78% said it improved their perception of the broker
- 94% would be comfortable using AI again
Notably, only 6% said they would have preferred to wait for business hours to speak to a human. Most valued immediate response over channel.
Economics
AI system cost: Approximately £1,200/month including usage Estimated annual value captured: £35,000-40,000 in new business alone ROI: Over 2,000%
And this excludes the value of better customer satisfaction, reduced callback workload, and improved claims information quality.
Lessons Learned
A year of operation taught several lessons.
Customers Adapt Quickly
Initial concern: would customers accept AI? In practice, customers adapted within seconds. When AI resolved their query immediately, they did not care whether they were speaking to human or machine.
Integration Is Everything
AI that cannot access systems is just a message-taker. The value came from AI reading policy data, writing claim records, and confirming information in real-time. Without integration, results would have been marginal.
Start Simple, Expand Gradually
Beginning with FNOL (a relatively structured process) built confidence. Expanding to information queries added value with lower risk. Quote capture came last because it was most complex. Phased deployment let the team learn and adjust.
Human Backup Matters
Some calls still need humans. Complex situations, emotional customers, unusual requests. The AI knowing when to escalate, and escalating smoothly, preserved customer experience for edge cases.
Competitors Noticed
Feedback from new customers: "I called three brokers at 7pm. You're the only one who answered." Availability became a competitive differentiator visible in conversion rates.
The New Normal
Today, this broker operates differently than competitors.
A customer has an accident at 11pm Saturday. They call their broker. AI answers immediately. Within three minutes, the claim is registered, the customer has a reference number, and they know what happens next. Monday morning, the claims team sees a complete FNOL ready for processing.
A prospect is comparing quotes Sunday afternoon. They call. AI captures their details and risk information. Monday morning, the sales team has a warm lead with complete information, ready to quote.
A policyholder needs their insurance certificate at 6:30am for a rental car pickup. They call. AI verifies identity and emails the certificate immediately.
This is not extraordinary service. It is just service, available when customers need it, not when the broker happens to be staffed.
The broker who never misses a call is not working harder. They are working smarter, using technology to be available around the clock without the cost of staffing around the clock.
And their competitors? Still checking voicemail Monday morning, calling back numbers that do not answer, wondering where the business went.
Ready to never miss a call?
SwiftCase Switchboard provides AI voice agents that answer every call, integrate with your systems, and handle customer queries 24/7. Your availability, transformed.
Book a demo | Learn about Switchboard | See the insurance solution