Problem
DeFi has almost 9,000 protocols across more than 500 blockchains, and each one works differently. A simple goal, like moving funds from lending on Aave (Ethereum) to a better yield on Solana, takes six actions across two networks: withdraw, swap, bridge, buy gas, swap again, deposit. Every step needs its own approval. Users jump between apps, and any step can fail, leaving funds stuck, delayed, or sent to the wrong network.
It's not only a beginner problem. We interviewed six experienced DeFi users in Brazil, and five had suffered exactly that. Interest is high (about half of US crypto exchange users wanted to try DeFi), yet 77% of Brazilians who know crypto say it is too complex. People want DeFi but lose time, and sometimes money, to human error.
Solution
FloFi lets users describe what they want in plain language, and turns it into one complete workflow.
1) Tell: "Move my Aave position to a better yield on Solana."
2) Build: FloFi finds the route across protocols and chains (swaps, bridges, lending, liquidity) and lays out every step.
3) Review: the user sees the exact steps, costs and route, and runs a simulation before anything happens.
4) Approve: the user signs with their own wallet. FloFi never holds keys or funds, and AI never has financial authority.
5) Execute and verify: FloFi runs the workflow, checks the on-chain result and shows a proof of what happened.
It's a real product, running on testnet today. The same engine works from our app, Telegram, AI chats like Claude and ChatGPT, and an API, so wallets and apps can offer multi-step DeFi without building it themselves.
Business model
Two revenue lines, both tied to completed workflows (planned; the app is free today).
1) B2B / API: wallets, apps and AI agents embed FloFi's engine. Partners pay per completed workflow, or we take a share of the fee they already charge, with a monthly minimum for larger partners. We start with paid pilots invoiced in fiat.
2) B2C: a small success fee on completed multi-step workflows, in the range of 0.05%-0.25% (about 0.1% in our base case), with a minimum and a maximum. No fee if a workflow fails. Simple swaps stay free, as in most wallets.
Ticket size: not measured yet. Our working hypothesis is US$100-1,000 per workflow, so about US$0.10-1 of fee at 0.1%. That is why costs per workflow and the B2B line matter most. Our beta will measure real tickets, frequency and cost per workflow before we fix prices.
Market
Global DeFi today (DefiLlama, Oct 2026): about $90B locked across thousands of protocols and 500+ chains, $304B in stablecoins, and about $8B of DEX volume per day. At that pace, that is roughly $2.5T a year.
On the people side, 741M people own crypto (Crypto.com, 2025), but only 40-70M are active on-chain users (a16z, 2025). That gap is the opportunity, and complexity is the barrier: 77% of Brazilians who know crypto call it too complex.
Our first market is the active users who already run multi-step DeFi, and the wallets and apps that serve them (MetaMask reports 30M+ monthly users, Phantom 15M). FloFi earns a small share of the value moved, or per use through the API, so even a small fraction of this activity is meaningful.
We have not yet measured our serviceable segment; our beta will give us the first real data.
Competitors
Today, DeFi users mostly rely on four kinds of tools:
1) Single-protocol tools: DeFi Saver (multi-step strategies and automation) and Glider (automated portfolios).
2) Wallet swaps, like MetaMask and Phantom: free and simple, but built for one trade at a time, not a full workflow across protocols and chains.
3) AI agents for DeFi, like HeyAnon: users type a goal and agents execute steps.
4) Developer infrastructure: Enso and Rhinestone sell multi-step execution to companies, not to end users.
Several earlier players (Furucombo, Zapper, Summer.fi, Brahma) have closed or been absorbed.
FloFi's difference: the AI proposes, but never has financial authority. Every workflow is simulated and reviewed in exact detail before signing, executed non-custodially with the user's own wallet, with recovery and a proof of every execution. One engine serves our app, AI chats, Telegram and API.
Competitive differentiation
Most DeFi AI assistants focus on individual commands, while execution protocols focus on routing transactions. FloFi brings both together in a unified workflow engine.
Users can build multi-step DeFi strategies through natural language or a visual canvas, both powered by the same underlying workflow representation. Unlike isolated swaps or bridges, FloFi is designed to orchestrate entire strategies across chains and protocols.
Our architecture combines pre-execution simulation, user-defined authorization limits, non-custodial execution, step-level recovery, and reconciliation.
The same engine is designed to power third-party applications through APIs and MCP, extending FloFi beyond a consumer interface into reusable DeFi infrastructure.
Our edge is not another AI chat for crypto. It's making complex, multichain financial workflows programmable, accessible, and executable from one unified experience.
Entry barrier
FloFi's barrier to entry is not the AI interface, but the execution infrastructure behind it.
Building a DeFi chatbot is relatively easy. Reliably orchestrating complex workflows across multiple blockchains, protocols, wallets, and failure scenarios is significantly harder.
FloFi combines a unified workflow engine, simulation, scoped authorization, execution recovery, and on-chain reconciliation. Each new chain and protocol requires specialized integrations, security engineering, and extensive testing.
As adoption grows, real execution data will help improve reliability, simulation accuracy, and failure recovery. Our API and MCP integrations can also create switching costs as third-party applications build on FloFi's infrastructure.
Our defensibility compounds through technical depth, proven execution reliability, and ecosystem integrations, rather than features competitors can easily replicate.
Traction
Not yet. Just 55 subscriptions waiting for test beta.