Scaling a Claims Portal Whose Pilot Results Cut Both Ways
A new digital claims intake portal was piloted for 10 weeks in two of eight regions. Results: 68% of claims submitted digitally against an 80% target; average handling time down 12%; first-contact resolution down from 74% to 66%; 31% of digital claims need rework because attachments are missing; call-centre volume down 9% but complaints up 18%. Regional managers describe the shortfall as customers not using the tool properly. The pilot excluded the broker channel, which accounts for 40% of claims in the participating regions. The vendor's dashboard counts submission attempts rather than accepted claims, and its summary reports the 80% target as met. The steering committee meets in three weeks to decide whether to scale to all eight regions. The pilot cost AUD 700,000; full rollout is estimated at AUD 2.4 million. The claims strategy states that any channel must not reduce first-contact resolution.
1. Evaluate the pilot results. Which measures are reliable, which are misleading, and what additional measures do you need before a scale decision?
Read the worked answer
Reliable enough to act on, with caveats: handling time, first-contact resolution, rework rate and complaints. These are outcome measures and they point in conflicting directions, which is the real finding.
Misleading as presented:
- The 80% target. The denominator excludes brokers at 40% of claims. If 68% applies to direct claims only, overall digital penetration is closer to 41%. The target may be met on one denominator and badly missed on the one the business cares about.
- The vendor's "submission attempts". Attempts are not accepted claims. This inflates performance and must be replaced with accepted, validated claims.
- Call volume down 9% alongside complaints up 18% suggests contacts may have shifted or become more complex, not disappeared.
Additional measures needed: cost per claim and cost per rework; time to resolution; complaint categories and severity; abandonment and error rates; digital penetration by customer segment and claim type; broker claim handling; staff workload and rework effort; and a comparison with non-pilot regions over the same period.
Also check comparability: a 10-week pilot may include a novelty effect and seasonal claim mix differences. Without a control comparison, the 12% handling-time gain is suggestive, not proven.
2. Diagnose the likely causes of the rework and the drop in first-contact resolution, and describe how you would isolate them.
Read the worked answer
Rework hypotheses: the form does not make required evidence clear; attachments are optional at submission or validated only downstream; file size or mobile capture limitations cause failed uploads; customers submit before they have the documents. These are design and validation issues, not customer carelessness - the managers' framing should be challenged with evidence.
First-contact resolution hypotheses: claims arrive incomplete or routed without the data needed to resolve them; digital claims may be counted differently in the FCR measure; staff are handling a mix of paper and digital with different procedures; complaints may reflect a longer or less visible journey.
How to isolate:
- Sample 100 rework cases and code the cause: missing document type, upload failure, illegible image, wrong claim type, customer error.
- Process-mine or trace claim paths for digital versus paper, comparing FCR by channel and claim type.
- Confirm the FCR definition and whether the baseline included broker claims.
- Segment by region, claim type and customer group; use a control chart to see whether the drop is sustained or a spike.
- Interview handlers and listen to a sample of complaint calls.
Most likely pattern to test: rework is a form and validation defect, and the FCR drop follows from incomplete submissions plus mixed procedures - both fixable, and both reasons not to scale unchanged.
3. What recommendation would you make to the steering committee, and how would you frame the enterprise-level considerations?
Read the worked answer
Recommendation: do not scale as-is. Use a conditional path: fix, extend, then scale through gates.
Proposed conditions before scaling:
- Redesign attachment capture with in-form validation and mandatory evidence by claim type; test with real customers.
- Correct the measurement set: accepted claims, agreed denominators including brokers, complaint categories, cost per claim.
- Extend the pilot for 8 weeks in the same regions plus one new region, with exit criteria such as rework below 10%, FCR at or above the 74% baseline, and digital penetration of at least 75% of direct claims.
- Design the broker channel before any enterprise rollout.
Enterprise-level framing for the committee:
- The claims strategy explicitly states that no channel may reduce first-contact resolution. Current results breach that condition, so scaling now would contradict an approved strategic constraint.
- Customer outcomes and complaints carry conduct and reputational risk, not just cost.
- The 2.4 million rollout should be released against demonstrated benefit, with a named benefit owner.
- Vendor reporting definitions need contractual correction, otherwise the committee will keep receiving an inflated view.
- Staffing, records retention and broker relationships need enterprise design, not regional workarounds.
Frame the decision as risk-managed sequencing rather than a binary scale or stop, and state what evidence would trigger full rollout.