Decisions that changed what we could promise

Our commercial property claims report once used final settlement time as its principal speed measure. That encouraged a complete view of each file, but it obscured cases where a substantial amount was already agreed while one smaller issue remained unresolved. A workshop could wait for cash to replace damaged equipment because a separate interruption calculation was still being examined. The file was progressing internally; the member could not act on that progress.

In 2023 I asked the claims director to establish a separate measure for the first undisputed payment. We limited the initial programme to defined commercial property cohorts and retained the final-settlement measure. Finance colleagues were concerned about reconciliation and recoverability of mistaken payments. Some experienced handlers worried that members would interpret a partial payment as agreement on the whole claim. Those objections shaped the process rather than being dismissed as resistance.

We created a short decision record identifying the accepted amount, unresolved elements and explanation given to the member. Experienced handlers received delegated authority within specified limits. Quality reviewers sampled paid and unpaid files, including cases in which a handler had decided against an interim payment. That last group mattered: reviewing only completed payments would have missed avoidable delay. We also tracked reopened issues and payment corrections so the new speed measure could not conceal a deterioration in accuracy.

By June 2025, median time to the first undisputed payment in eligible claims had fallen from 23 to 9 days. We did not describe that result as the settlement time for every commercial claim. The programme required A$1.1 million of additional training and review expenditure in its first year. It gave members earlier access to established amounts without pretending that unresolved questions had disappeared. I would now involve broker service teams sooner. Their early explanations sometimes differed from the wording claims colleagues had carefully tested.

During the 2022 renewal cycle, Brindleharbour Mutual received 420 formal complaints about pricing in a reviewed commercial portfolio. Members were often told that market conditions explained the increase. That phrase combined several different causes and gave frontline staff little help when a member asked why a neighbouring business had received a different offer. The absence of a usable explanation also encouraged exceptions whose combined cost was difficult to see.

I commissioned a joint pricing and distribution review. We separated five components: expected claims, repair-cost assumptions, catastrophe protection, portfolio experience and individual exposure changes. We did not promise that every member could reproduce the actuarial calculation. We required a defensible account of the main movement and a clear route for correcting factual errors. The commercial team argued that more detailed explanations could create longer calls and expose inconsistencies in our own data. They were right about both risks.

We trained a first group of advisers with actual anonymised renewal cases and corrected the explanation where it broke down. A proposed concession had to identify its funding effect and the reason it was appropriate for that member. We also accepted that some business should leave if its desired price could not be supported. The board reviewed retention alongside the quality of the remaining portfolio, rather than treating a lower retention number as an automatic management failure.

Across two annual cycles, formal complaints in the reviewed portfolio fell from 420 to 255. Retention moved from 91% to 88%, and we recorded the lost premium rather than presenting the programme as an uncomplicated success. The important change was a more honest renewal conversation and a visible record of exceptions. I would have tested the first explanatory letters with a broader range of small businesses; our initial language assumed more familiarity with insurance cost components than several members had.

Before the 2024 catastrophe-protection renewal, I asked how confidently we could assemble evidence for a large recovery. The purchasing paper compared capacity and price in detail. It said less about the location records and claims documentation that would support a submission after a widespread event. A sample of relevant historical files found missing location-level evidence in 18% of cases. The omission did not mean that every recovery would fail, but it made our stated readiness difficult to defend.

The finance director preferred to complete the data correction before changing the protection purchase. The reinsurance team was concerned that waiting would leave too little time to negotiate terms. I asked both teams to distinguish what could be corrected before renewal from what would remain an operating dependency. We took the board a combined decision: additional work on records and a protection structure better aligned with the exposures we intended to retain, at A$4.2 million more annual cost.

Claims specialists joined the renewal discussions and tested the proposed evidence requirements against sample files. We assigned owners to missing fields, created an exception register and required a documented recovery rehearsal using a synthetic event. The exercise exposed two hand-offs that were assumed rather than assigned. We corrected those responsibilities before describing the process as ready. The additional premium reduced funds available for other discretionary projects, including a planned distribution enhancement.

The subsequent file sample showed missing evidence in 4% of reviewed cases. I treat that as a control improvement, not evidence that a future recovery will be simple. Contract interpretation and event facts can still produce difficult discussions. The enterprise gained a shared understanding of what it had bought and what it needed to maintain. My principal lesson was to involve claims colleagues at the start of the buying process, before commercial options had already narrowed around a preferred proposal.

In 2021, our supplier register suggested that we had several alternatives for urgent commercial property repairs. A closer review found that some firms depended on the same subcontractors and specialist equipment. Different contract names did not necessarily represent independent capacity. The problem became visible during a regional loss simulation in which several member premises needed similar work at the same time.

Procurement proposed adding more panel firms. Claims leaders preferred retaining a smaller group whose quality they knew. I asked for a dependency map and a trial that tested both positions. We identified the work for which genuine alternatives existed and the work constrained by scarce specialist capability. We then selected two additional regional firms for a limited pilot, retained inspections and paid for readiness arrangements where a credible capacity commitment could be established.

The financial trade-off was explicit. The pilot increased annual supplier-management and standby expenditure by A$780,000. We also accepted a smaller procurement saving than the original tender objective. In return, we required more transparent allocation rules and evidence of actual availability during a surge. Member communications explained when an interim measure or payment might be more practical than waiting for our preferred repair route. We avoided presenting cash as a universal substitute for a managed repair.

Within the pilot portfolio, median time from authorisation to repair commencement fell from 19 to 12 days over twelve months. Repeat visits for defective work remained at 3.2%, which meant the improvement had not yet solved every quality concern. We retained that qualification in the board report. The programme changed how we assessed supplier resilience: subcontractor dependencies and observed delivery mattered more than panel size. I would have included smaller rural jobs earlier, because travel and minimum-job economics created constraints that the metropolitan sample understated.

As eastern regional director at Kestervane Underwriting, I inherited a niche property segment that contributed A$42 million of annual gross premium. The reported growth was attractive, but renewal files contained an increasing number of pricing and wording exceptions. Each exception had a plausible commercial explanation. Together they described a product different from the one our underwriting authority and operating capability were designed to support.

In 2012 I asked for a cohort review covering claims experience, referral volumes and the cost of servicing exceptions. The distribution head argued that we should invest in a revised product because brokers valued our responsiveness. The underwriting team believed that a narrower appetite would be safer. I proposed a staged withdrawal from the least coherent part of the segment while retaining clearly defined risks that met the intended terms. This sacrificed volume and required uncomfortable broker conversations.

We stopped new business in the affected cohort, honoured existing commitments and gave brokers advance notice of renewal decisions. A small transition team checked that communications matched the actual policy position. We did not use the withdrawal as a reason to rush open claims or impose retrospective conditions. I moved two experienced colleagues into broker support for the transition and accepted that their usual growth objectives would need revision.

The affected premium reduced by A$17 million over eighteen months. Referral volumes in the remaining segment fell by 36%, and the team could explain the retained appetite without relying on informal workarounds. I did not attribute subsequent claims movements entirely to the withdrawal; weather and claim development also affected the result. The decision taught me that commercial relevance requires a product we can administer consistently. If repeating it, I would establish the cohort review earlier, before the accumulation of exceptions had created expectations that were costly for both brokers and us to unwind.