--- title: "We Ran a Market Simulation On Our Own Product – Here’s What 19 AI Agents Predicted About visibility.so" url: https://sanjayshankar.me/ai-market-simulation-visibility-so-validation/ description: "We ran 19 AI agents through a market simulation for visibility.so. Here's what they predicted about pricing, adoption, and objections - and what we're building next." author: Sanjay Shankar date_published: 2026-06-15T13:22:17+00:00 date_modified: 2026-09-10T10:13:57+00:00 categories: [Product, SaaS] tags: ["Agentic SEO", "AI Market Simulation", "Build in Public", "GEO", "mirofish", "Product Validation", "visibility.so"] reading_time_minutes: 6 --- # We Ran a Market Simulation On Our Own Product – Here’s What 19 AI Agents Predicted About visibility.so ![Visibility.so using Mirofish](https://sanjayshankar.me/wp-content/uploads/2026/06/visibility-mirofish\1.webp) Most startups wait until launch to learn if their market is ready. We’re doing it differently. Before we build the next major version of [visibility.so](https://visibility.so), we ran a full market simulation on our own product. Nineteen AI agents — each role-playing a different market persona — debated visibility.so across 50 rounds on simulated Twitter and Reddit. We wanted to know: - Are SEO agencies actually ready to hand over work to AI agents? - What’s the real objection hiding beneath “I’ll think about it”? - Does our all-in-one pricing actually compete with Semrush and Ahrefs? - Which customer segment should we go after first? Here’s what the simulation told us — and how it’s shaping what we build next. > **What this post is:** A transparent look at how we’re using AI to validate our product roadmap — before we write the code. > > > > **What this post isn’t:** A claim that we’ve solved everything. We haven’t. But we now know exactly what to solve — and that’s the most valuable thing a simulation can give you. --- ## The Setup: 19 Agents, 50 Rounds, 2 Platforms We used [MiroFish](https://github.com/666ghj/MiroFish) — an open-source AI market simulation engine — to run the experiment. If you’ve read my [MiroFish setup guide](https://sanjayshankar.me/mirofish-setup-guide-ai-market-simulation/) or [how-to guide](https://sanjayshankar.me/mirofish-ai-how-to-use-guide/), you know how this works: spawn AI agents with distinct personas, motivations, and biases, drop them into a simulated social environment, and watch them debate your product. **Our agent lineup for this simulation:** | Persona Type | Count | What They Represent | | --- | --- | --- | | SEO Agency Owner | 5 | 5-50 clients, consolidating tool stacks, margin pressure | | SaaS Founder | 4 | 10-200 person companies, doing their own SEO | | Indie Hacker / Solopreneur | 4 | Solo founders, bootstrapped, budget-sensitive | | Enterprise Buyer | 3 | VPs at 500+ person companies, compliance-focused | | Competitor (Semrush, Ahrefs, Moz, Screaming Frog) | 3 | Market incumbents reacting to a new entrant | Each agent was given a detailed persona — background, goals, budget, pain points — and let loose on a simulated social feed for 50 rounds across Twitter and Reddit. The result: **136 actions** (posts, replies, debates, and reactions). ![Visibility.so Mirofish Setup](https://sanjayshankar.me/wp-content/uploads/2026/06/visibility-mirofish-setup\1.webp) --- ## 3 Things the Simulation Confirmed ### 1. Agencies Are Ready to Consolidate — And Our All-In-One Pricing Is Perfectly Timed The simulation repeatedly surfaced a pattern: agencies are tired of managing multiple SEO subscriptions. > *“Agencies are consolidating tool stacks to reduce costs, including potentially for visibility.so.”* This matched what we’ve heard from beta users. But seeing it independently emerge from 19 AI agents — who knew nothing about each other — was strong validation. **The pricing dynamic the simulation picked up on:** | Tool | Monthly Cost | What’s Extra | | --- | --- | --- | | [Semrush](https://semrush.com) | $200+/mo | DataForSEO credits + LLM API keys | | [Ahrefs](https://ahrefs.com) | $99+/mo | DataForSEO credits + LLM API keys | | **[visibility.so](https://visibility.so)** | **$99-199/mo** | **Nothing — all included** | > *“visibility.so is a flat fee, while Semrush costs $200+/mo and requires extra payment for DataForSEO credits and user’s own LLM API keys.”* ### 2. “AI Agents Doing SEO Work” Is the Real Differentiator Every competitor tool in the simulation — Semrush, Ahrefs, Moz, Screaming Frog — was characterized as “analytics-only.” None of them executes SEO work. The agents consistently noted that visibility.so’s model — where AI agents actually *run* audits, *track* rankings, *monitor* AI visibility, and *surface* content opportunities — is a fundamentally different category than traditional SEO tools. > *“visibility.so is a competitor to Screaming Frog, Semrush, Ahrefs, and Moz.”* But here’s the nuance the simulation revealed: **we’re not competing on features.** Semrush has more reports. Ahrefs has a bigger backlink index. We compete on *outcome* – getting SEO work done without hiring more people. The simulation confirmed that agencies feel this pain acutely. ### 3. Small Agencies and Indie Founders Will Adopt First The simulation was clear on adoption order: **First wave:** SEO agencies with 5-50 clients who feel the margin squeeze **Second wave:** SaaS founders doing their own SEO, wanting automation **Third wave:** Enterprise teams (after more validation and case studies) > *“Agencies often juggle multiple clients and need tools that are not only effective but also straightforward to set up and use without much overhead.”* This matches our beta data perfectly. --- ## 2 Objections We Need to Address The simulation wasn’t all validation. It surfaced real concerns that we’re taking seriously as we build the next version. ### Objection 1: “Will this integrate with my existing workflow?” > *“If visibility.so requires extensive training or doesn’t integrate well with my current strategies, it could end up wasting both time and resources.”* This is the #1 concern that came up across multiple personas. Switching SEO tools is painful. Nobody wants to spend a week configuring a new platform only to find it doesn’t play well with their stack. ### Objection 2: “Is this actually useful, or is it AI hype?” > *“I’ve been burned by expensive tools that my team never used, skeptical of AI-powered claims after trying several duds.”* This one hit close to home. There’s a lot of “AI-washing” in the SEO tools space right now, and the simulation agents — especially the indie founders — were rightfully skeptical. --- ## What We’re Building Next Here’s the honest part. We haven’t solved these objections yet. But the simulation gave us a clear, prioritised list of what to work on. ### 1. Smoother Onboarding (Addressing the Integration Concern) The simulation confirmed that our onboarding needs to be faster and more guided. We’re building a setup wizard that takes you from signup to your first agent-run audit in under 10 minutes – including connecting GSC, GA, and choosing your first project. ### 2. Better Communication About How Agents Work > *“Advocating for transparent communication from the developers regarding methodologies and data sources would ensure our decisions are well-informed.”* Trust is earned through transparency. We’re working on an “agent explainer” feature that shows exactly what data each agent pulled, which decisions it made, and why — so you’re never wondering if an AI agent is making things up. ### 3. Pre-Built Templates for Common SEO Tasks > *“As a solo founder with a limited budget, I’m focused on finding tools that offer free trials or affordable pricing without overwhelming complexity during setup.”* Small teams shouldn’t need an SEO degree to get started. We’re building pre-configured agent templates for common workflows: weekly site audits, competitor monitoring, content gap analysis, and GEO tracking. One click, and they’re running. ### 4. Community and Social Proof > *“Engaging with current users about their experiences could provide valuable insights into its advantages and pitfalls.”* We’re expanding our [Discord community](https://discord.gg/5ySTt7KRuQ) with dedicated channels for sharing agent configurations, audit templates, and real results from beta users. Nothing builds trust like hearing from someone who’s already using the product. --- ## What’s Coming in 1-2 Weeks We’re shipping the next major version of visibility.so in **1-2 weeks**, and these simulation insights are directly shaping what goes into it. **Here’s a sneak peek:** - Simplified onboarding with guided setup - Agent transparency logs — see what your agents are doing, in plain English - One-click audit templates for common SEO tasks - Expanded integrations If you want to see how this plays out, **[start a 7-day free trial](https://app.visibility.so)** now, and you’ll get the new features the moment they ship. Early users are grandfathered at their join rate. --- ## Why We Ran This Simulation We ran this because we believe in building what the market actually needs – not what we *think* the market needs. MiroFish cost us pennies. The insights it generated – about pricing perception, adoption order, and hidden objections – saved us weeks of guesswork. More importantly, it gave us confidence to double down on our roadmap and the clarity to fix what actually matters. If you’re building a product, I’d encourage you to try the same. [My MiroFish guides](https://sanjayshankar.me/mirofish-setup-guide-ai-market-simulation/) walk through the entire process. As for visibility.so – the simulation confirmed our direction. And in 1-2 weeks, you’ll see the result. --- ## Frequently Asked Questions **What is MiroFish?** MiroFish is an open-source AI market simulation engine that spawns AI agents with distinct personas and runs them through simulated social environments. Free on [GitHub](https://github.com/666ghj/MiroFish). I wrote a [setup guide](https://sanjayshankar.me/mirofish-setup-guide-ai-market-simulation/) and a [how-to guide](https://sanjayshankar.me/mirofish-ai-how-to-use-guide/). **How many agents did you use?** 19 agents across 5 persona types: SEO agency owners, SaaS founders, indie hackers, enterprise buyers, and competitor tools. **Are these real people?** No. Each agent is an AI language model (GPT-4o-mini / GPT-4o) given a detailed persona — background, goals, budget, pain points — and asked to role-play that character on simulated social feeds. **How accurate are the results?** Directional, not predictive. Excellent for surfacing blind spots and validating assumptions. Not a replacement for customer conversations. **Did the simulation change your roadmap?** It confirmed our direction and helped us prioritise. The onboarding improvements and transparency features were already on our list – the simulation moved them to the top. **What's coming in the next version?** Simplified onboarding, agent transparency logs, one-click audit templates, and expanded integrations. Shipping in 1-2 weeks. **Can I try visibility.so now?** Yes. [Start a 7-day free trial](https://app.visibility.so) — no credit card required. You’ll get the new features as soon as they ship, grandfathered at your join rate. --- *Found this useful? Share it with a founder who’s debating their next feature. And if you run your own simulation, tag me on [X (@sanjayshankarr)](https://x.com/sanjayshankarr) — I’d love to hear what you learned.* *Up next in 1-2 weeks: “We shipped the features our simulation told us to build — here’s what changed.”*