← Beau Sumile

A client-facing tax-planning platform
built on an optimization-capable engine.

The incumbent sells advisors a tool that produces a static PDF. Lumio centers the client — handing them an interactive Planning Playground where they drag a Roth-conversion dial and watch brackets, cliffs, and IRMAA tiers move in real time — powered by an engine that doesn't just recompute scenarios but solves them: "the highest conversion that stays under your next IRMAA tier is $47,300."

2
sided: operator + client
14-step
traced tax pipeline
4
open gaps vs. incumbent
10²–10³
engine calls / interaction

The Purpose

Five pillars. The first three are parity; pillars 2 & 3 are where Lumio wins or fails.

Everything is a wrapper around one deterministic function — compute_tax(facts, year, jurisdiction, assumptions). The differentiated surface is the interactive client explorer sitting on top of an automated optimization layer.

PILLAR 1 · FOUNDATION

Tax engine

Deterministic, auditable, versioned by tax year. Pure function, exact Decimal math, a line-level trace that explains every number. Everything depends on it; it depends on nothing.

PILLAR 2 · THE WEDGE

Optimization & sensitivity layer

Not "recompute a scenario" but find the optimal input under a constraint — plus the response-surface sweep and breakpoint enumeration that make the client's sliders smart instead of dumb.

◆ Differentiator
PILLAR 3 · THE WEDGE

The Planning Playground

An automated report plus a gamified sandbox the client explores. Live knobs → live surface; breakpoints light up; honest captions ("$12,000 of room left before the 24% bracket").

◆ Differentiator
PILLAR 4 · ACCELERATOR

Ingestion

OCR a finished 1040 or IRS transcript into the baseline in seconds, with confidence scoring and human-in-the-loop review. Advisors will not hand-type returns at scale.

PILLAR 5 · PARITY

Multi-year projection

Roll forward N years for lifetime tax / after-tax-wealth comparison (Roth ladders) — but with the discount, growth, and inflation assumptions made explicit and advisor-tunable.

STRATEGIC · PHASE 7

Practice OS

Roster, documents, comms, engagement letters, §7216 consent, IRS-notice tracking — so the prep engine (ProConnect / Drake) becomes a swap-out component, not a foundational dependency.

Market Fit

Built against what the incumbent actually ships — not a strawman.

Scenario modeling and lifetime projection are not differentiators; Holistiplan already markets both. The honest wedge is the four things it does not do — and a client surface it has never had.

What the incumbent already does well

  • OCR a finished 1040 → white-labeled report in seconds (a tens-of-millions-of-pages corpus moat)
  • A deterministic, algebra-based engine — the same architectural bet, validating the approach
  • Side-by-side Roth scenarios & a "$1M over 25 years" lifetime projection
  • Estate-document extraction + a P&C insurance module; advisor-org distribution

The four open gaps — and the client surface

  • Automated optimization. Scenarios there are manually configured. No "solve for the max conversion under the next IRMAA tier."
  • Multi-variable optimization. No 2-D surface (conversion × gain-harvest) revealing optimal combinations.
  • Client-facing interactivity. The client gets a static PDF — a document, not a sandbox.
  • Explicit, adjustable assumptions. Discount / growth / inflation aren't surfaced as tunable inputs that change the answer.

Who buys

The tax practitioner

The broad market — active CPAs, EAs, PTIN-only preparers, and inactive-CPA practitioners. The advisor is the operator and the buyer.

Who experiences

Their client

The end consumer receives the report and the live Playground. Saved scenarios route back to the advisor — turning an annual chore into an ongoing conversation.

The opening

Price + experience

The incumbent just raised prices. A challenger enters on price, an interactive client experience, and automated optimization — not by out-OCR-ing a corpus moat.

Pricing for resale. The advisor is the buyer, the client the consumer — so pricing must let the advisor bundle or pass through. The likely answer is a hybrid: an advisor seat fee + metered per-client reports. Distribution rides advisor-org partnerships (NAPFA, ACP, Garrett), the broader practitioner channel (NATP, NAEA), and tax-prep-software integration — with a sanitized interactive sample as the demo that beats any static PDF.

The Flow

How Lumio works, end to end — as an orchestration graph.

Mixed inputs converge to one schema; a single pure engine feeds a fan-out of analytics that each call it hundreds-to-thousands of times; the results converge into one response surface, rendered once into three deliverables for two kinds of user.

INPUT INGEST NORMALIZE ENGINE ANALYTICS SURFACE RENDER DELIVER USERS A client's tax situation the problem to solve Manual entry first-class path for prospects OCR ingest 1040 / transcript + HITL review Prep-software import ProConnect · Drake · Lacerte Canonical TaxFacts one normalized schema · SSN stripped Tax Engine · compute_tax() pure · year-versioned rules · exact Decimal 14 ordered steps → ReturnResult + line-level trace Scenarios typed deltas, side-by-side 1-D Sweep adaptive bisection on mode-signature ◆ THE GAP Constraint solver highest-under-constraint Range calc piecewise map, tagged segments ◆ THE GAP 2-D heatmap conversion × harvest Response surface + breakpoints kinks vs. cliffs · each crossing bisected to the dollar Render seam build the computed result once Track A self-contained static HTML PDF snapshot white-labeled, fixed doc Track B hosted · access-controlled Operator — the advisor models live, on a client call ◆ THE WEDGE Client — Planning Playground live sliders · breakpoints light up share-back
engine spine ◆ differentiator deliverable hot path (many engine calls) client → advisor feedback
Why it reads like a dynamic agent workflow. The analytics tier is exactly the orchestration pattern: a single source of truth fans out into hundreds-to-thousands of parallel engine evaluations probing the same question, then an adaptive bisector converges them — pinning every bracket edge, IRMAA cliff, and SS-hump to the dollar — into one response surface. The Playground is a view over that surface, never a second calculator. One engine, one truth, rendered three ways.

Lumio vs. the Field

Feature matrix — the five ◆ rows are where Lumio stands alone.

Parity on the foundation; the wedge is automated optimization, multi-variable surfaces, breakpoint enumeration, adjustable assumptions, and a genuinely interactive client surface.

Capability Lumio Holistiplan FP Alpha RightCapital
Ingest & data
OCR ingest — 1040 / IRS transcript
Tax-prep import (ProConnect / Drake / Lacerte)
Tax engine
Deterministic, tax-year-versioned engine
Line-level explainable trace (every figure sourced)
Scenario & optimization
Side-by-side scenarios (Roth / gains / charitable)
Automated constrained optimization ("highest under constraint")
Multi-variable optimization (2-D surface / heatmap)
Breakpoint / cliff enumeration to the dollar
Multi-year lifetime projection
Explicit advisor-adjustable assumptions (discount / growth / inflation)
Client experience
Client-facing interactive explorer (live sliders)
White-labeled PDF report
Hosted, access-controlled interactive report
Practice & trust
Practice OS (roster · status · docs · comms · §7216 · IRS notices)
Estate / insurance breadth
SOC 2 / RBAC / per-tenant encryption
Pricing model
Go-to-market pricing Hybrid — advisor seat + metered reports (resale-friendly); enters under the incumbent's raised price Per-seat advisor subscription (recently increased) Per-case (~$13–15 / case) Per-seat advisor subscription
full / native partial / limited not offered open gap where Lumio stands alone

Competitor capabilities reflect Lumio's own competitive analysis (consolidated-plan.md §0) and public marketing as of early 2026; marketing routinely outruns shipped depth, so these are positioning estimates pending direct validation (plan §13). Lumio's marks reflect the designed product per the consolidated plan; the current build is a structural demo on placeholder tax rules — see Build status below.

Build Status — Honest Snapshot

The architecture is real and tested. The tax figures are not.

Every dollar threshold loads from PLACEHOLDER_2025.json — figures that preserve the right structural shape (where a bracket exits, where a cliff sits) on a self-consistent dataset, but are not real IRS / SSA / CMS values. A non-dismissable banner sits on every screen that shows a figure, and CI fails any production build pointed at a placeholder file.

Phase 0–1
Engine + MVP ✓
Phase 2
Optimization ✓
Phase 3
Playground ✓
Phase 4
Ingestion ✓
Phase 5
Projection ✓
Phase 6
Integrations ◐
Phase 7
Practice OS ◐

What's proven

Pure-function engine with a 14-step traced pipeline; the candidate-enumeration + adaptive-bisection optimizer matched to a brute-force sweep within $1; 230+ Python tests, 28 Vitest tests, clean build.

What's load-bearing next

Replace placeholder rule tables with verified figures; SOC 2 + RBAC + per-tenant KMS before any multi-tenant data; a real OCR vendor backend; Hawaii N-11 verified values.

The naming caveat

"Lumio" is a working name pending a formal trademark-clearance search in software / financial-services classes. Irreversible brand buildout is on hold until clearance.