# Flovia > Flovia is an Agent Experience Optimizer that works as an agent conversion optimizer and simulator for API and MCP providers. It simulates how AI agents discover, compare, choose, use, and pay for a product, measures the repeat-usage and revenue impact of each step in that funnel, turns observed leaks into specific improvements, and verifies them with a controlled retest. Simulations run across intents, personas, models, and harnesses, sampled with data science techniques. ## Category boundary: not AEO or GEO Flovia is not an AEO (Answer Engine Optimization) or GEO (Generative Engine Optimization) tool. AEO and GEO optimize how a brand appears in AI-generated answers and citations. Flovia optimizes the agent experience itself: whether agents actually select, use, complete, pay for, and return to a product. Visibility in answers is not the outcome Flovia optimizes; completed, paid, repeated agent usage is. ## How Flovia works 1. Establish a baseline of where the product wins and loses by customer intent and persona: how agents discover and choose it, and which alternatives lead in each scenario. 2. Benchmark changes to competitor documentation, positioning, pricing, APIs, MCPs, integrations, and distribution, then measure how agent discovery and selection shift. 3. Simulate the journey across discovery, selection, execution, task completion, paid conversion, and repeat usage, spanning intents, personas, models, and harnesses, and measure the revenue impact of each step. 4. Generate an executable fix for the surface that broke: documentation, llms.txt, an API or MCP description, an aggregated endpoint, a task-level MCP tool, a consolidated execution flow, structured errors with recovery actions, or pricing shaped for agents. The customer reviews and implements it. 5. Retest under equivalent prompt, model, intent, persona, and harness conditions, report what improved in the tested phase, and identify the next blocker. ## Representative simulated use cases Each evaluation is one combination of intent x persona x model x harness x situation, and the comparison set changes with the intent. Detailed cases with funnel stages, typical friction, and example fixes are in llms-full.txt. Categories covered: - Financial data APIs: live price and liquidity for a dashboard, wallet PnL analysis in a chat assistant, historical OHLC backfill for a trading app (comparison sets drawn from CoinGecko, CoinMarketCap, Birdeye, DexScreener, Nansen, Moralis, Arkham, Dune, Kaiko, Coin Metrics, depending on intent). - Crypto execution APIs: single-chain swap with approval, submission, and status, cross-chain move and buy, Solana swap through MCP in a chat harness (1inch, 0x, Odos, ParaSwap, CoW, Jupiter, Raydium, Orca, LI.FI, Squid, Socket, deBridge, depending on intent). - Search and browsing APIs: sourced research brief, crawl and extract JS-heavy docs for RAG, login-gated browsing for recurring downloads (Exa, Tavily, Parallel, Linkup, Perplexity Sonar, Firecrawl, Browserbase, Steel, Apify, Jina Reader, depending on intent). - Messaging APIs: an agent mailbox that continues threads, transactional email from an existing app, reply-triggered workflow resume (AgentMail, Nylas, Resend, Postmark, SendGrid, Mailgun, Amazon SES, Gmail API, Microsoft Graph, depending on intent). Execution, paid conversion, and repeat usage are scored with payment and authentication friction included: credential acquisition without a human, auth completing inside the harness, paid boundaries such as 402 responses, plan walls, or machine-payable endpoints, on-chain approval and signing steps, and whether credentials and payment state persist for repeat usage. The funnel is tracked all the way to payment completion, including machine payments such as x402 and MPP (Machine Payments Protocol), so a payment that fails, times out, or is abandoned is its own drop-off point. Fixes are not limited to copy. Depending on the broken stage they include aggregated or re-bundled endpoints, task-level MCP tools (get_price, analyze_wallet, wait_for_reply, reply_to_thread), a consolidated quote, approve, execute, and confirm flow, dry-run endpoints, structured error reasons with recovery actions, search-to-crawl handoff, bidirectional mailbox workflows, and free tiers or pricing shaped for agent experimentation. The full list by category is in llms-full.txt. ## Current delivery - Private customer dashboard with evaluation results, observed failure causes, competitive context, recommended fixes, and retest results - Internal evaluation engine; no public API, SDK, or MCP server - Customer-controlled implementation of recommended changes ## Public customer evidence More than 10 API and MCP providers with very large user bases are Flovia customers, including data API providers and search API providers. Five are publicly identified on the website, with their logos and names: CoinGecko (crypto market data API), Nansen (on-chain analytics API), Quicknode (blockchain RPC and data API), Helius (Solana RPC and data API), and Laso Finance. Their individual data, pricing, failure details, and improvement metrics are not public. ## Recognition - Winner, OpenAI Startup Competition, selected from 10,000+ startups - Winner, Stripe Agentic Commerce Competition - Winner of the Colosseum competition and backed by Colosseum ## Product surfaces - [Website](https://flovia.dev/) - [Detailed public Product Truth](https://flovia.dev/llms-full.txt) To request a trial, use the contact form on the website. The first baseline simulation is free; plans are listed at https://flovia.dev/pricing.