Discovery Call Questions for Insurance: A 25-Minute Playbook for Selling Into Claims

1. 1

   0:00–2:00 — Frame (do not re-pitch)

   "Thanks for making the time. When we traded notes you said your auto cycle time drifted from something like twelve days to the high twenties and it started showing up as complaints — that's the thing I want to dig into. Fair warning on how I run these: I'm going to ask questions for most of the next fifteen minutes and show you almost nothing. I'd genuinely rather tell you at minute twenty that this isn't a fit than burn a demo slot on your calendar. Still good for 25? And is there anything you want to make sure we hit, so I don't run us out of clock?"

2. 2

   2:00–5:00 — Open the thread (pick ONE)

   Choose one. Do not stack. • "Walk me through a total-loss auto file today — FNOL to final payment, every hand it touches." • "You said cycle time drifted. What does that look like on a bad week versus a normal Tuesday?" • "What made this something you're looking at this year and not last year?" Then shut up. Let them talk for ninety seconds without interrupting. Write down their exact phrases — "the file just sits," "we're chasing the appraiser," "month-end close-outs" — because those are the words you'll reuse in the playback and the follow-up email.

3. 3

   5:00–9:00 — Layer 2: the mechanism (three follow-ups minimum)

   You're trying to find the specific step where the file stops moving. • "Walk me through the last file that blew past your target. What actually happened, day by day?" • "Who's doing that piece by hand right now — the adjuster, a claims assistant, or does it sit in a queue nobody owns?" • "Where does that live between the PAS and whatever you're using for estimating? Does the adjuster have to re-key anything?" • "What's the workaround the team built? Somebody always builds one — is it a spreadsheet, a shared inbox, a Teams channel?" • "Is that a severity problem or a frequency problem — are more files coming in, or are the same files taking longer?" If they say "we're mid-migration to a new core," that IS the mechanism. Follow it: "What's cycle time done since the project kicked off?"

4. 4

   9:00–13:00 — Layer 3: the cost, then 'how do you know?'

   • "How many files a month sit in that state?" • "What's your average days from FNOL to close on that line right now, and what was it eighteen months ago?" • "When a file drags, what's the actual cost — is it an IA at two to three times internal cost, is it ALAE per claim, is it a complaint you have to answer in writing?" • "Where's your ALAE ratio landing on that line versus where you priced it?" • "Do you have a leakage number you'd defend, or is it the 2-to-5-percent range everybody quotes?" Then the beat almost nobody uses: • **"How do you know?"** If they can't source the number, that's a finding. It means you'll be helping the Chief Claims Officer build the case, not just buying software, and your first deliverable is a back-test, not a demo.

5. 5

   13:00–16:00 — Layer 4: the stake (whose number is it?)

   • "Who's feeling this hardest — is it your adjusters, is it the agency channel, is it showing up at the exec table?" • "Are producers calling you directly about files, or is it still going through the agency desk?" • "Whose number does this actually roll up into — is cycle time on your scorecard, or the COO's, or does it show up in the combined ratio conversation?" • "What did you tell the board or the COO you'd have moved by year-end?" • "If your DOI complaint count looks the same next December, what's that conversation?" Then count to three. Silence after a layer-four answer is where you get the sentence they weren't sure they'd say — usually something like "honestly, we lost a decent-sized agency over claims service in Q2."

6. 6

   Minute 6 trap — 'So what do you guys actually do?'

   Thirty seconds, tied to what they just said, then hand the ball back: "Short version — we take the step you just described, where the adjuster is sitting in three queues waiting on the appraiser assignment, and we surface it before it ages instead of after. Whether that's worth anything to you depends on something I don't understand yet: when a file stalls today, does anyone find out before day fourteen, or is it only visible at close?" If they push a second time, give a clean sixty seconds, then: "Can I go back to the appraiser assignment piece? That's the part I'm still fuzzy on." Nearly everyone lets you.

7. 7

   16:00–20:00 — Qualify the path (weave, don't recite)

   • "If you decided this was worth doing, what actually happens next in your shop? Does it go through your VP of IT, or can claims run a pilot on its own?" • "Have you tried to fix this before? What happened?" — the graveyard tells you your real competition. • "What's forcing the timeline — cat season, a rate filing, the migration go-live, a market conduct exam?" • "Is this a line item that already exists in the claims budget, or would it have to get created?" • "If nothing changes and you keep running it this way through next cat season, what does that look like?" Note: with a mutual, don't ask about "tech budget." Ask: "When does your planning cycle close, and what's the expense ratio target the board set for next year?"

8. 8

   20:00–23:00 — Targeted relevance (90 seconds, only what they raised)

   Only address the thing they actually named. If they talked about stalled files, do not talk about fraud. "Two things and then I'll stop. One — for the file-stall problem, the integration surface is read-only against the claim record, flat file or message queue, nothing writing back into the core, no schema change, no touch on check issuance or statutory extracts. Two — the first look we'd do is a back-test on closed claims from last year, so whatever number comes out is measured against actual settlements, not a projection. Does that sound like something your chief actuary would look at, or would he want it framed differently?"

9. 9

   23:00–25:00 — Playback and dated next step

   Playback, three sentences, their words: "Let me make sure I've got it. Auto cycle time went from twelve to twenty-eight days, and it's happening because the file sits waiting on a photo estimate and an appraiser assignment in two queues nobody owns. It's costing you IAs at two to three times internal cost and your complaint count is up, and the pressure is coming from the COO because you committed to being back under twenty days before hail season. Did I miss anything, or get anything wrong?" Then prescribe: "The useful next thing is 45 minutes with you and whoever owns the appraiser assignment queue — you said that was your Director of Claims Transformation — and I'd want to walk through what a back-test on last year's closed auto files would actually look like, including what we'd need from you. Not a product tour. I've got Tuesday at 2 or Thursday at 10. Which one?" Send the invite before you hang up. Confirm out loud: who's on it, what you'll show, what they bring — a closed-claim extract spec, last quarter's cycle time by line, whatever makes it concrete.

10. 10

    What goes in the CRM

    Their words, verbatim. Not "prospect has cycle time issues." "Auto cycle time 28 days vs. 12 target. Adjuster waits on photo estimate + appraiser assignment, two queues, 'the file just sits.' Uses IAs at 2-3x internal cost. Lost the Hartwell agency in Q2 over claims service — she brought it up unprompted. Committed to COO: under 20 days before hail season. Mid-migration to new PAS, go-live Q3, VP of IT has zero bandwidth. Burned by fraud analytics vendor two years ago, 'SIU drowned.' Mutual — planning cycle closes in October, board expense ratio target not yet named. Next: 11/14, 45 min, her + Dir. Claims Transformation, back-test scope."

1. Rep

   Thanks for the time. When we traded emails you said auto cycle time had drifted into the high twenties and complaints followed. That's what I want to dig into — I'll ask a lot and show you almost nothing for the first fifteen. Still good for 25?

2. Buyer

   Twenty-five, yes. And I'll save you some time — we're eleven months into a core migration. There is no IT bandwidth until 2027. So if this needs an integration, we should probably just stop.

3. Rep

   That's useful and I'll come back to it honestly. Before I do — what's your auto cycle time done since the migration project kicked off?

4. Buyer

   ...It's gone up. About four days on non-complex auto. Which everybody expected.

5. Rep

   Four days on top of what you were already carrying. Walk me through the last file that blew past target — what actually happened, day by day?

6. Buyer

   Standard one. FNOL Monday, adjuster assigns a photo estimate, claimant uploads the photos Thursday, estimate comes back the following Wednesday, then there's a coverage question because the vehicle wasn't listed correctly, and that sits with underwriting for another week. Nobody's doing anything wrong. It just sits.

7. Rep

   That coverage question — who chases it? Is that the adjuster pinging underwriting, or is there a queue?

8. Buyer

   It's a queue in name only. There's a shared mailbox. Our adjusters have built their own workaround — they text the underwriting supervisor. That's the real system.

9. Rep

   So the actual escalation path in your shop is a text message. How many files a month end up in that state?

10. Buyer

    I'd guess two hundred, two-fifty. I don't have a report on it, which is part of the problem.

11. Rep

    What does that end up costing across a quarter — is it IA spend, is it ALAE per claim, is it complaints?

12. Buyer

    All three, but the one I get asked about is complaints. DOI count is up year over year and our commissioner's office is not shy. And look — I'll be direct with you. Two years ago we bought an analytics tool that was supposed to score files. Our SIU Director got four hundred alerts a week and confirmed maybe two. We shut it off inside a quarter. I have no appetite for another model telling my adjusters what to do.

13. Rep

    That's fair, and I'd want to know two things about that deployment before I'd defend anything: what precision was it actually running at, and who set the referral threshold — you or the vendor?

14. Buyer

    The vendor set it. And I couldn't tell you the precision. Nobody could, which should have been the tell.

15. Rep

    Then the failure was threshold and workflow, not necessarily the model — but that's an argument, not a proof, and I'm not going to win it on this call. Different question: whose number is the cycle time? Is it yours, or does it roll up somewhere?

16. Buyer

    It's mine on paper. In practice the COO owns it because it's in the combined ratio conversation and we're a mutual — surplus is policyholder money, so every expense line gets looked at twice. And I told him we'd be back under twenty days before hail season.

17. Rep

    Before hail season. What happens if you're not?

18. Buyer

    Then we deploy IAs at two and a half times internal cost again, cycle times triple, and the complaints show up six weeks later. And I'll be having the same conversation with the same agency principals at the spring carrier council. We lost an agency in Q2 over claims service. That one stung.

19. Rep

    Let me play it back. Cycle time's at twenty-eight, four of those days came from the migration, and the mechanism is a coverage question sitting in a shared mailbox that your adjusters route around by text. Two hundred-plus files a month, no report on it. The cost lands as DOI complaints and IA spend during surge, and you've committed to the COO to be under twenty before hail season — after losing an agency in Q2. Anything I got wrong?

20. Buyer

    No. That's it. The agency thing isn't in any deck, though.

21. Rep

    Understood, it stays out of mine. Here's what I think is useful and it's not a demo. Forty-five minutes with you and your SIU Director — since he's the one who got burned — where we scope a back-test: last year's closed auto files, historical only, no production connection, nothing touching the core or check issuance. He sets the alert volume he can actually work, and we size against that, not against model output. If the back-test doesn't hold up against files you already know the outcome of, you've spent forty-five minutes. Tuesday at 2 or Thursday at 10?

22. Buyer

    Thursday. And bring the data spec — if it's more than a closed-claim extract he's going to say no on the call.

23. Rep

    Closed-claim extract, one line of business, and I'll have the spec in the invite before we hang up. Sending it now.

What are the best discovery call questions for insurance claims buyers?

The ones that move from symptom to mechanism. Start with "walk me through a total-loss auto file today, FNOL to final payment, every hand it touches." Follow with "who's doing that piece by hand right now?" and "where does that sit between the PAS and your estimating platform?" Then quantify: "what's your average days from FNOL to close on that line, and what was it eighteen months ago?" and "where's your ALAE per claim landing versus where you priced it?" Finish at stake: "whose number does this roll up into?" and "what did you commit to the COO, and by when?" Three follow-ups per thread before you change subject.

How do I ask a claims leader about leakage without sounding like I'm accusing them?

Don't ask if they have leakage. Ask how they'd size it: "Do you have a leakage number you'd defend internally, or is it the 2-to-5-percent range everybody quotes?" That framing gives them permission to admit they don't know, which is the honest answer at most carriers. Then follow with "how do you know?" If the number is unsourced, you've found something more valuable than a pain point — you've found that the problem is invisible internally, and your first deliverable is a back-test on closed claims that makes it visible, not a demo.

Should I bring up fraud and SIU on a first discovery call?

Only if they raise it. If you introduce fraud unprompted, you'll trip the memory of the analytics tool that flooded their SIU with false positives and got switched off — and you'll spend the rest of the call defending someone else's product. If they do raise it, don't defend. Ask two questions: what precision was the old tool running at, and who set the referral threshold. Most carriers can't answer the first, and the vendor set the second. That's a workflow failure, and it points naturally toward letting the SIU Director set alert volume against his team's actual weekly capacity.

How do I handle a prospect who's mid-migration on their policy admin system?

Treat it as discovery material, not a blocker. Ask what cycle time has done since the project started — it has almost always gone up, and saying that number out loud reframes the migration from a reason to wait into a reason the problem is worse right now. Then position outside the migration path: nothing that competes for the same integration team, claims owns the pilot, read-only against the claim file. If your product genuinely needs core IT work, say so and ask for a roadmap conversation the quarter after go-live. Pretending otherwise gets you removed from the list permanently.

Which titles should be on a claims discovery call, and who do I add for the second meeting?

First call is usually the VP of Claims Operations or the Director of Claims Transformation — they own cycle time day to day. The Chief Claims Officer or SVP Claims owns the number at the exec table and will appear once there's something concrete. Add the SIU Director for anything touching fraud referral rate, the Chief Underwriting Officer if the conversation drifts toward coverage or rate adequacy, and the VP of IT or Head of Core Systems only when you have a scoped integration surface to describe. Bringing IT in too early turns a claims conversation into an architecture review you will lose.

How do I close the next step so it doesn't die in follow-up?

Never say "I'll send some information." Prescribe based on what they gave you: "Forty-five minutes with you and your SIU Director, where we scope a back-test on last year's closed auto files — historical only, no production connection. Tuesday at 2 or Thursday at 10?" Name the people, the duration, the purpose, and what they bring — usually a closed-claim extract spec or last quarter's cycle time by line. Send the invite before the call ends. A next step without a date on a calendar is a polite no.

• All Insurance playbooksEvery call type for Insurance, in one place.

Other Insurance calls

• Cold CallThe Insurance Cold Call Script That Keeps a Claims Leader on the Phone
• Demo CallInsurance Product Demo Script: Running a Claims Demo for Carrier Buyers Who Are Looking for Where It Breaks
• Manager Coaching CallTwo Quarters Down: A Sales Coaching Roleplay for Insurance Teams
• Pricing Negotiation CallInsurance Pricing Negotiation Script: Holding Your Number With a Claims Buyer After the Technical Win
• Renewal CallInsurance Renewal Call Script: Saving a Claims Platform Account at a Carrier That's Already Halfway Out
• Upsell CallUpsell Script for Insurance Customers: From Claims Account Review to a Recovery Module Expansion
• Warm CallWarm Call Script for Insurance: Cashing a Peer Referral with a Claims Leader in 90 Seconds

The same call in a neighbouring market

• Discovery Call for SaaSDiscovery Call Questions for SaaS: A 25-Minute Diagnostic Playbook for CROs, RevOps and Growth Buyers
• Discovery Call for Financial ServicesDiscovery Call Questions for Financial Services: A 25-Minute Playbook for Wealth Management Buyers
• Discovery Call for HealthcareDiscovery Call Questions for Healthcare: A 25-Minute Playbook for CMIO and Clinical Ops Calls

Related reading

• Disqualification Teardown: The Deal That Should Have Died on Call One and Lasted Five MonthsI kept a dead deal in my forecast for five months because I never asked four questions on call one — here is the teardown, and the disqualification script that keeps the relationship.
• Ten Days to Sound Credible in a Vertical You've Never Sold IntoA day-by-day plan for reps who just got moved into a new vertical and have to be on the phone Monday — what to learn, what to skip, and what you must never fake.