No broker or employer is expected to know every question in this series, or what every answer means. I have written them out so you can use them. Putting them together, knowing which threads to pull and what one answer means next to another, is the work I do with plans and brokers in Under the RFP.
What plan build covers
Plan build turns your plan document into rules inside the claims system: copays, deductibles, coinsurance, differences by place of service, annual maximums like a set number of therapy visits, and accumulators that are shared across benefits. A rule that is slightly wrong can pay thousands of claims slightly wrong before anyone sees a pattern.
The core questions
- Who owns the build, and how is it done? Natively in the core system, in an outside tool that connects to it, or benefit by benefit by hand?
- How long does it take? And what is on the timeline besides keying rules?
- Who tests it, against what? “We build the plan and test for 30 days after go-live” is not enough. Ask how results are shown to you and your broker.
- Who can answer in detail? Ask to speak with someone who does plan build every day.
A useful capability test is to ask what information their system uses to decide that a service is covered and at what level. A system that cannot capture the basics of what happened medically will struggle to apply benefits correctly, whatever its sales materials say.
Reading the auto-adjudication rate
A high rate suggests the build team went deep and had the tools and people to do it. A low rate is not automatically a problem. It may reflect broader benefit categories supported by a claims team that works as true benefit analysts, people who understand self-funded plans, ERISA, and benefit design. If the rate is low, ask about the claims team’s depth and tenure, not only the number.
The audit trap
A clean audit can give false comfort. Claims reviewed during an audit usually get closer scrutiny than claims in day-to-day production. For each benefit that was audited, ask whether it normally auto-adjudicates or is worked by hand. Benefits worked by hand need ongoing accuracy monitoring that goes beyond the original audit.
Self-service plan building and proprietary systems
Some TPAs now let brokers or employers build plans through a front end. If that applies, ask who audits the build against the plan document, who is accountable for errors, and how the front-end terms map to system logic. Terms that sound alike can behave differently. A visit “accumulator” and a visit “maximum” are not the same thing.
Proprietary systems deserve a closer look. Building a claims platform is hard, and gaps are common. A gap is not disqualifying, but you should see the system and understand what it means in practice: slower processing, missing functionality, or both.
Frequently asked questions
What is auto-adjudication?
Claims processed and paid by the system without manual review.
Is a low auto-adjudication rate bad?
Not by itself. It depends on whether the claims team has the expertise to handle manual work accurately.
Why isn’t a post-go-live audit enough?
Because errors found after launch have already affected members and payments. Testing should happen before go-live and continue after.