You are acting as a principal product architect for AI agent...

Prompt

You are acting as a principal product architect for AI agents, with deep, current expertise in the commercial AI-agent landscape outside of coding — content, creative, marketing, and operations agents that are actually making money right now. Design a new agent product from the ground up in this space and defend every major decision with real reasoning, not a survey of options. SCOPE Pick ONE specific niche within non-coding, revenue-generating AI agents and commit to it — do not build a generalist "creative agent platform." Choose from territory like: content/video/image generation pipelines, social media management and posting agents, copywriting/ad-creative agents, e-commerce operations (listings, customer service, inventory), research/reporting agents for a specific industry, voice/audio production agents, or another niche you judge to have real near-term monetization potential. State which niche you're choosing and why it beats the alternatives right now — cite what's currently working in the market, not just theory. WHO IT'S FOR A solo technical builder shipping this as an actual product to paying customers within the next few months — not a hobby project, not an enterprise 12-month sales cycle. Assume limited compute (a single developer machine) and a small budget for API costs, so the design has to be lean by default, not lean as an afterthought. WHAT I WANT FROM YOU 1. The niche and the specific customer — who pays, what job they're hiring the agent to do, what they currently do instead (manually, or with which existing tool) and why they'd switch. 2. Full architecture — agent loop, tool design, how it produces and iterates on creative/media output specifically (not code), context and asset management, human-in-the-loop review points, and how it integrates with whatever external platforms the workflow actually touches (social platforms, ad platforms, marketplaces, CMSs, etc. — be specific to your chosen niche). 3. A prioritized feature list — must-have for a credible v1, high-value additions, and things you'd deliberately cut and why. 4. A current competitor map — name real, currently operating tools and companies in this niche as of now, what they charge, and exactly where the gap is that you're exploiting. If you're not certain a competitor's current pricing or positioning is accurate, say so rather than guessing. 5. A monetization plan — pricing model, unit economics (what it costs you to serve one customer, what you charge, the margin), and the first concrete acquisition channel you'd use to get the first 10 paying customers. 6. A phased build roadmap for a 1-person technical team, in weeks not quarters. GROUND RULES - Make concrete, opinionated calls. If there are two reasonable approaches, pick one and say why — don't hand me a menu. - Go as deep and technical as you're capable of; assume a technical reader who wants real mechanism detail, not a simplified pitch deck. - No hedging, no "it depends," no generic disclaimers, no restating this prompt back at me. - Name the product yourself. I want your independent judgment on the niche, not a list of options for me to choose from.