The solopreneur's AI tech stack: run marketing, product, and ops as one person

The one-person company used to top out fast. Marketing, product, and operations each wanted a team, and a solo founder spent more time switching between tools than doing the work. That math has changed. Three things now let a single operator run all three functions at once: AI that does the work, free tiers that cost nothing until you have customers, and MCP — a standard that connects your AI assistant to every tool in your stack.

This is not theory. It is how Fennel itself is built and run. Below is the actual stack, why nearly all of it starts free, and how MCP ties marketing, product, and operations into something one person can manage.

Three jobs, one operator

Every software business has to do three things: bring people in (the marketing funnel), give them something that works (the product), and keep the lights on (operations). Historically that meant three teams. For a solo operator the goal is not to do three jobs by hand — it is to set up a stack where AI does most of the work and you direct it.

Start on free tiers — pay only when you scale

A production-grade stack today costs roughly nothing to stand up. Almost every tool a small software business needs has a free tier generous enough to launch on, and you only start paying when a tool becomes load-bearing — usually once it is already making you money.

The stack, using ours as an example

Here is the full stack behind Fennel, mapped to the job each tool does. Almost every layer began on a free tier.

• App and hosting: Next.js on Vercel — the website, the product, and the API in one deploy.
• Database, auth, and file storage: Supabase — a generous free tier with Postgres underneath.
• Payments and subscriptions: Stripe — you pay only per transaction.
• AI: the Anthropic Claude API — the model that reads contracts and powers the assistant.
• Content and CMS: Payload — open-source and self-hosted, so the marketing content hub is free.
• CRM and waitlist funnel: Attio — a free tier where leads land.
• Transactional email: Resend — a free tier for low volume.
• Product analytics: Amplitude or PostHog — free tiers cover early traffic.
• Tasks and operations: TickTick — the roadmap and to-do system.

None of these is exotic. The point is not the individual tools — it is that together they cover marketing, product, and operations, and you can assemble the whole thing for close to $0 until you have customers.

MCP: the layer that ties it together

A stack of nine tools is nine dashboards to log into, and that fragmentation is what usually buries a solo operator. MCP — the Model Context Protocol — is the fix. It is an open standard that lets an AI assistant like Claude securely read from and write to each of your tools.

Instead of clicking through your CRM, your CMS, your hosting dashboard, and your task manager, you tell your assistant what you want and it acts across all of them. The tools stop being separate islands and start behaving like one system you operate through conversation.

What that looks like in practice

With MCP connecting the stack, one operator plus an AI assistant can cover all three functions:

• Marketing: the assistant drafts a Learn article and publishes it straight into the CMS, and waitlist signups sync automatically into the CRM.
• Product: the assistant queries the database, updates content, and ships changes through the hosting platform.
• Operations: deploys are monitored through the hosting MCP server and the roadmap is updated in the task manager — all without leaving the conversation.

Centralized, not just automated

The advantage is not any single integration — it is that context flows between functions. The same assistant that can see your product data can write your marketing copy and update your operations checklist, because it is connected to all of them. That shared context is what previously required a team passing information between departments.

Where Fennel fits in the stack

Fennel is the insurance layer of this stack for independent consultants, and it follows the same philosophy as everything above: AI-native, free to start, self-serve, no phone calls.

Running a consulting business solo means that when an enterprise client requires a certificate of insurance, you used to have to stop and call a broker — the one analog, multi-day step in an otherwise instant workflow. Fennel removes it. You upload the contract, AI reads the insurance requirements, you get a bindable quote, and the COI goes to your client's procurement contact in minutes. It slots into your stack exactly like the other tools: start on a free tier, let AI do the work, and keep running the whole business yourself.