Pitch deck · July 2026
Felix
fēlīx, -īcis (Latin) — fortunate; favoured by fortune.
A private wealth manager who answers only to you. Felix sees everything you own, thinks about it every night, and hands you clear, explained moves each morning — and never touches your money.
01 · The problem
People with real, growing portfolios are managing them on gut feel — because every form of help asks them to give something up.
Human advisors
$5,000–$20,000 a year on a $1M portfolio, slow to reach, and often paid to steer you toward their own products.
Robo-advisors
Take custody, drop you into a generic risk bucket, and charge a percentage of everything you own — forever.
Doing it yourself
Spreadsheets, Reddit threads, and a CPA who calls after the tax year is over. Nobody sees the whole picture.
Meanwhile the money itself is scattered — brokerages, funds, retirement accounts, cash, property — with no single, current view of the whole. The big decisions get made blind. And the quiet truth of the industry: most of what a $20K/year advisor actually does is run a known playbook. That playbook is automatable.
02 · The market
The target sits in a dead zone the industry has never priced for: too much complexity for a robo, not enough appetite to hand a human 1% of everything, every year.
US households, age 25–35, with $500K–$5M investable — growing on equity comp and the great wealth transfer.
what those households already spend on financial advice — advisor fees, robo fees, tax prep.
transferring from Boomers to younger generations through 2045. They won't use their parents' advisor.
The pricing map has a hole in it: DIY planning tools cluster at $100–$200 a year, managed money starts at $2,500 and runs to $10,000+. The $200–$2,000/year band — ongoing, comprehensive, advisory-only — is where the category is forming, and the first entrants are already repricing into it. Beyond the US: Felix is built jurisdiction-parameterized from day one, with India — one of the fastest-growing retail-investor bases in the world — as the second market.
03 · Why now
- AI crossed the advisor-grade line. Explaining a portfolio the way a great advisor would — grounded, personal, in plain English — became feasible in the last two years. The analysis itself was always automatable; the conversation wasn't.
- Self-directed investing surged, sophistication didn't. A generation opened brokerage accounts without a plan. Accumulation has outpaced advice.
- Fee awareness is at an all-time high. The target user knows exactly what 1% AUM compounds to. They pay for software; they resent tolls on their net worth.
- The category is being validated in real time. In the three months to July 2026, the closest comparable repriced from ~$125 to $299–$1,499/year and registered as an SEC fiduciary; a YC-backed entrant launched conversational advice at $15/month. The window is open — and visibly narrowing.
04 · The product
Connect everything once. Every night Felix re-prices the whole picture, recomputes the analysis, and decides what matters. Every morning you get a briefing and a short list of explained moves.
One complete picture
Every account in a single, daily-updated view of your net worth — stocks, funds, retirement, cash, plus anything added by hand.
Advisor-grade analysis
Look-through exposure, drift against your plan, concentration, real performance, and the fees you're quietly paying — recomputed continuously by tested code.
Actions, not homework
Rebalancing, tax-loss harvesting, cash deployment — each a structured recommendation with the reasoning laid out. You act in your own accounts; Felix verifies you actually did.
Ask anything
A chat that already knows your portfolio, across every screen — with an incognito mode for the questions you'd rather not save.
News that's about you
Market and company news distilled to what actually moves your holdings — surfaced before it matters, not after.
The plan at the center
Onboarding builds your model portfolio the way a planner would. Everything else — drift, actions, advice — is measured against it.
05 · A morning with Felix
“Markets were flat yesterday; your portfolio slipped 0.3%, led by your semiconductor exposure — you hold more of it than you think, through two overlapping funds. You've drifted 4.2% from your plan's equity target. Three things worth your attention today: a rebalancing trade that brings you back inside your band, a tax-loss harvest worth roughly $1,900 if done before the wash-sale window closes, and $12,000 in cash that's been idle for forty days. Each comes with the numbers it was computed from.”
You accept the harvest, dismiss the rest, and place one trade at your own broker. Tomorrow night, Felix's verifier matches the imported transaction against the accepted action — the loop closes with evidence, not assumption. That loop is the retention engine: the product is new every morning because the market is.
06 · AI you can audit
The design principle that makes advice trustworthy: deterministic facts, generated words.
- The LLM never does math. Drift, tax lots, wash-sale windows, retirement projections — every number comes from tested, deterministic code. The AI reads those facts; it cannot invent them.
- Every recommendation shows its work. Structured, schema-validated output that cites the exact facts it came from — with validators enforcing that the quantities match the deterministic plans they cite.
- Execution is impossible by construction. Felix holds read-only credentials at every broker. There is no trade scope to misuse — the advisory-only promise is enforced by architecture, not by a prompt.
One incorrect tax calculation destroys trust forever; a black-box recommendation never earns it. This architecture exists because accuracy and explainability are existential in this category.
07 · Business model
A flat subscription. No percentage of assets, no custody, no products to push — the incentives point the same way the advice does.
Priced like software
Research points to $10–$30/month at entry, expanding toward $50–$150/month as the full playbook lands — against the $5K–$20K/year a human charges for the same moves.
Near-zero marginal cost
The expensive AI work — digesting news, scoring, summarizing — is a function of the content, not the reader. It runs once and fans out; a marginal user costs roughly $20/month, chat included.
No conflict to outgrow
AUM businesses can't build this without cannibalizing themselves. A subscription business never faces the choice.
08 · Competition
Map the field on two axes — how much it covers, and who keeps control — and the industry sorts into three clusters, none of them here:
They manage your money
Wealthfront, Betterment, Empower, Arta. Comprehensive but custodial — you hand over the keys and pay AUM forever. Their business model is the moat around Felix, not around them.
They watch, but don't advise
Kubera, Monarch, Copilot, Empower's free dashboard. You keep control, but they're trackers — no playbook, no actions, no accountability.
They advise, thinly
Mezzi ($299–$1,499/yr, newly SEC-registered), Astor ($15/mo), PortfolioPilot — plus general AI assistants doing free, session-based approximations. The quadrant is contested but unowned.
Nobody yet combines the full advisor playbook — rebalancing, tax optimization, asset location, retirement — with an advisory-only model and analysis you can audit. Felix's bet is that the winner of this quadrant is decided by depth and trust: deterministic analytics, a daily loop, and verified actions, not another chat window over Plaid data.
09 · Moats
- Incumbents structurally can't follow. Every funded competitor with the data to do this monetizes AUM. Advisory-only software that says “keep your money where it is” cannibalizes their core revenue.
- The playbook's edge cases are the barrier. Wash sales across accounts, cross-account asset location, jurisdiction-specific tax law. The core algorithm is a weekend; advisor-grade correctness is years of edge cases.
- Switching costs compound. A canonical, reconciled history of everything you own — plus every action taken and verified against it — gets harder to walk away from every month.
- Trust accrues to the auditable. In a field about to fill with confident black boxes, “every number traceable to tested code” is a durable brand position.
10 · Where it stands
Felix is past the paper stage. Discovery research validated the white space; the full system is designed end-to-end and published openly; the build is underway as a friends-and-family deployment — deliberately fee-free, which keeps it outside the heavy advisory regimes in both target countries while the loop is proven with real portfolios.
Designed end-to-end
Data rails (SnapTrade for the US, CAS for India), a canonical store, deterministic analytics, the daily agent, chat, and tenancy — the complete design lives in the Field Notes.
Absurdly lean
Built solo. The entire deployment — data, AI, and infrastructure — runs at roughly $85–95/month. The experiment costs almost nothing to run honestly.
Sequenced to learn
Holdings visible first, then computed facts, then the daily briefing, verification, news, chat. Retention — the one existential question — gets tested with every phase.
The thinking is public: the Field Notes hold the research, architecture, and every decision — including the ones that got reversed.
11 · The vision
Private wealth management has always existed — for the people who need it least. Felix's bet is that the advisor's playbook is software, the conversation is finally solvable, and control never needed to be the price of good advice.
The end state: everyone with a real portfolio — in the US, in India, anywhere — has a wealth manager that works only for them. Fortune, favoured.