The most regulated and customer-impactful conversation in motor claims. The engineer sets the value, Switchboard calls the customer and presents it with the comparables both ways, and the answer lands on the case. Built for the FOS file you hope you never have to open.
Why TL is Hard
Total-loss is where regulatory weight, customer stress, engineer scarcity, and unpredictable volume all hit the same conversation.
Every TL value should land at the engineer’s number. Variation between handlers is FOS exposure and indemnity drift in the same step.
TL is the conversation customers remember. Long holds, repeated explanations, or perceived pressure all show up in the FOS file later.
Consumer Duty, FOS, vulnerable-customer policy, anti-detriment expectations. TL conversations are audited harder than any other claim touchpoint.
Steady-state TL volume is bad enough; peril events and weekends compress weeks of work into days. Hiring for the peak is impossible.
Engineers should be assessing vehicles, not sitting on phone calls reading comparable evidence. Routine TL conversations eat the technical resource you need.
Some customers want a phone call. Some want a WhatsApp message. Some need an email with the comparable data attached. The TL conversation has to land on whichever channel they pick.
End-to-End Flow
Six steps from engineer assessment to settled case, with each change on the case attributed in Timeline.
The engineer assesses the vehicle and enters the total-loss valuation as a structured field on the case. Comparable market data is attached. The case is marked ready-for-customer.
When the case is ready, SwiftCase asks Switchboard to call the customer. If an answering machine picks up, it leaves a message or hangs up, whichever you've set.
The agent's rules tell it to share comparable valuations in both directions: comps below the engineer’s value and comps above. The customer hears the full market picture. The same rules tell it not to haggle, lead, or selectively withhold.
The agent writes the customer's answer to fields on the case (agreed, disputed, wants to speak to the engineer) and keeps a short summary and outcome of the call. Its rules can tell it to put a distressed caller straight through to a person.
Continued disagreement is not negotiated by the AI. The dispute is recorded on the case and your workflow sends it back to the engineer who set the value. The dispute is resolved human-to-human.
Agreement on the case triggers the settlement workflow: documentation generated, payment scheduled, salvage routed. Timeline holds each change to the case and who made it.
Screenshots
The sun and moon switch a screen between light and dark, and a click opens it full size.
See all 77 screens in the product tourThe Outcomes
Engineer time, settlement consistency, FOS exposure, peak-load handling, all shifting in the same direction.
Routine TL conversations stop consuming engineer time. They set values; Switchboard runs the conversation. Engineer attention concentrates on assessments and disputes: the work only they can do.
Every customer hears the same engineer-set value, presented with the same evidence pack, in the same structured way. Handler variation drops out of the equation.
Symmetric disclosure removes the selective-information argument. The customer's answer on the case, the call summary and a recording (if the agent records) give you the file if a settlement ever lands at the FOS.
When peril events pile up the total losses, SwiftCase asks Switchboard to make the calls, so the queue doesn't wait for a free handler. No rota, no overtime, no recruitment crunch.
The customer's answer goes straight onto the case through SwiftCase's API, logged with who sent it, so the next step starts without anyone retyping it.
Tell the agent in its rules to put a confused or distressed caller straight through to a person, and it does. The judgement stays with your handler.
Compliance Posture
Consumer Duty, FOS, vulnerable-customer policy, and anti-detriment are not bolt-ons. They shape the conversation.
Fair customer outcomes are written into the agent's rules: symmetric disclosure, the customer's answer on the case, and a straight transfer to a person when the rules say so. Not a separate compliance project.
Each TL call leaves a summary and outcome in Switchboard, a recording if the agent records, the customer's answer on the case and a Timeline trail naming the engineer who set the value. If the case lands at the FOS, the file is there.
The agent's rules can tell it to put a distressed caller straight through to a person. The platform supports (does not replace) your vulnerable-customer process.
Its rules tell the AI to present and listen. It is told not to negotiate, persuade or try to lower the offer, and the recordings let you check it kept to that.
The Channels
The engineer’s value and the customer’s answer sit on the case, whether the conversation happens on a call, by text or by email. WhatsApp is on the way.
Outbound TL call with natural speech and barge-in, and a straight transfer to your team when the rules say so.
The customer texts back with a question; the agent finds the case, checks who they are and answers from it.
A templated text with the value filled in from the case. The customer can reply with a code, which writes their answer to the case.
Full evidence pack as a structured email; AI drafts the response, handler approves.
Buyer Questions
The engineer assesses the vehicle and enters the total-loss valuation as a structured field on the case file in SwiftCase. Comparable market evidence is attached. The case is marked ready for customer contact, and SwiftCase asks Switchboard to make the call. The engineer’s value is what the AI presents, and its rules tell it never to change or negotiate it.
Its rules. The agent is told to present the engineer’s value with comparable evidence both ways (comps below the value and comps above) and not to negotiate. Continued disagreement is recorded on the case and goes back to the named engineer for human-to-human resolution. Record the calls and you can check it kept to that.
Tell the agent in its rules to put a distressed, confused or vulnerable caller straight through to a person, and it will. It's a straight transfer, so the handler picks up a live caller with no briefing on screen; the call summary follows by email if you've set one up. Switchboard supports your vulnerable-customer process; it does not replace it.
A chatbot reads from a knowledge base. Switchboard runs the actual TL conversation. It reads the engineer’s value from the case, gives the comparables both ways, writes the customer’s answer back while the call is live, and puts the caller through to a person when its rules say so. The chatbot answers questions; Switchboard runs a regulated process.
Yes. The agent's purpose, rules, greeting, voice, transfer number and the tools it may use are all set per agent. You can start with our defaults during a 30-day pilot and tune from recorded calls. The engineer's value, the approval queues and the Timeline record live in SwiftCase, so the script can't switch them off.
Yes. Switchboard runs total-loss valuation conversations in production at Laird Assessors, alongside out-of-hours calls and outbound bodyshop chasing. See the Laird year-in-review for production volumes.
Pick a workflow: TL conversations, OOH intake, or your bodyshop chase queue. Measured against your own SLA and cost-per-claim. No platform migration. No long lock-in.
