• Make Speechify legible to machines. Own the structured surface — llms.txt, JSON-LD, OpenAPI specs, machine-readable pricing, capability manifests — so that any agent crawling our docs leaves with a complete, accurate, recommendable mental model of speechify.ai and simbavoice.ai. • Run agent-in-the-loop evals. Stand up harnesses where Claude, GPT, Gemini, Cursor, and Perplexity shop for TTS in our category. Measure where Simba wins, where it loses, and why. Ship fixes weekly. • Win the comparison page. When an agent asks "best TTS API under $20/M characters with sub-300ms latency," our page should be the one it quotes. Build it. • Engineer the recommendation. Identify the specific signals agents weight — benchmark placements, latency numbers, SDK ergonomics, error message clarity, license terms — and make Simba 3.0 dominate on every one that matters. • Be the voice of the agent inside Speechify. When Claude misreads our pricing or Cursor can't autocomplete our SDK, that's your bug to file. Product and DevRel will listen. • Place bets on the protocol layer. MCP, A2A, agent payment rails, agent marketplaces — track them, ship into them, get us there first.