⚡ Case Study · Internal Product

Killing 20-minute quote times — I built an in-house agent quoting portal that returns the best plan in under 30 seconds.

2025 · Design + Build + Deploy · National health insurance brokerage

The Challenge

Brokerage agents were burning real time on every single quote. The plan data lived across multiple carrier PDFs, rate sheets, and pricing tables — each formatted differently, each requiring the agent to manually cross-reference age bands, coverage tiers (Employee, Employee + Spouse, Employee + Children, Family), networks (PHCS, ANTHEM, others), and supplement add-ons (dental, vision, critical illness, accident).

A typical client conversation looked like this: client gives ages, household structure, and state on the phone. The agent puts them on hold, opens 3–4 carrier rate documents, finds the right age band for each member, cross-references the coverage tier, hand-calculates supplement pricing, then assembles a recommendation. Average quote time: 15–25 minutes per client. Error rate on manual lookups: meaningful enough to cost deals when a wrong number got quoted.

Worse, the team was scaling. Adding agents meant onboarding each one on a manual lookup process — and every additional carrier file added permutations to the lookup matrix. The system was about to collapse under its own weight.

The Solution

Designed, built, and deployed a React-based internal quoting portal that ingests the brokerage's pricing data and returns ranked plan recommendations in seconds. The agent enters household composition once — primary applicant age + gender + state, optional spouse, optional dependents — and the portal does everything else.

Built the system around four core engines working together: a tier-resolution engine that automatically determines coverage tier (Employee / +Spouse / +Children / Family) from household composition; an age-band lookup that maps each member to the correct rating bucket (18–29, 30–44, 45–54, 55–64) per plan; a supplement pricing layer that applies per-member dental, vision, critical illness, and accident add-ons; and a deduplication + ranking layer that sorts results by best value within the agent's specified budget ceiling.

Architected the data layer to be CSV-driven so the operations team can update pricing without engineering involvement — new carrier plans, rate updates, and supplement changes are dropped into the spreadsheet and the portal picks them up. Built UHC indemnity plans as a separate code path (individual rather than household-tiered pricing). Added a cart system so agents can assemble multi-plan packages with running subtotals before quoting.

"What used to take twenty minutes now takes thirty seconds. Our newer agents quote like veterans because the portal handles everything they'd otherwise need years of muscle memory for."
— Operations Lead, National Health Insurance Brokerage

The Impact

Average time-to-quote dropped from 15–25 minutes to under 30 seconds. Agents stopped putting clients on hold to dig through PDFs. New-hire ramp time on quoting collapsed from weeks to a single training session — the portal replaced the institutional knowledge that used to live in senior agents' heads.

Quote accuracy went up too: with manual lookups eliminated, the wrong-number errors that used to surface on policy issuance disappeared. The operations team now controls pricing through a CSV they update directly — no engineering bottleneck when carriers change rates.

Beyond the speed wins, the portal became a repeatable asset. The same architecture is being extended to additional carriers and product lines as the brokerage grows.

30s
Quote Time
40×
Faster vs Manual
200+
Plan Combos Indexed
0
Manual PDFs Needed

Time per client quote — pre vs. post portal

Average minutes from intake to plan recommendation
25 min 15 min 5 min 0 ~20 min Manual lookup (before) 30 sec Portal (after) 40× faster

How It Works (Under the Hood)

The portal is a single-page React app deployed alongside the brokerage's internal tools. Architecture decisions worth calling out:

  • CSV-driven data layer: ops uploads carrier plan files in a standard format; portal parses on load. Zero engineering needed for rate updates.
  • Tier auto-resolution: household composition → coverage tier is fully derived. Agent never picks the tier manually; the math just runs.
  • Multi-member age band lookup: each household member is rated against their own age band before household pricing assembles. Older children get correct supplement pricing.
  • Special-case carrier logic: UHC indemnity is priced per-member (not per-household), with a dedicated code path that branches automatically based on carrier.
  • Cart system: agents can assemble multi-plan packages (e.g., medical + dental + accident) and see a running subtotal before generating the quote.
  • State exclusions: jurisdiction-specific rules (e.g., Louisiana supplement exclusions) handled in the pricing logic, not as agent training material.

The takeaway: the highest-leverage paid media work isn't always paid media. Sometimes the biggest revenue lift is fixing the 5 minutes after a lead becomes a quote — which is exactly what this portal does. Marketing leadership at six-figure-per-month ad spend means thinking past the platform and into the operations.

Need this for your funnel?

I help in-house teams and agency clients build paid-media systems that actually convert intent into revenue.