Tools change. Principles don’t. What follows is the stack I’m running as of early 2026, organized by function rather than by vendor. I’ve included what I’d change if I were starting from scratch, because what I use now reflects decisions I made two years ago that I wouldn’t necessarily make again.
Total infrastructure cost for a solo operator: $200 to $500 per month. Compare that to a single additional hire at $60k to $80k per year.
A Note on Team Size and Tool Selection
Before I share my stack, one thing I want to call out because it matters for how you read this section.
I had the luxury of building inside Copy.ai for three years. That meant I had access to a workflow-building platform purpose-built for chaining AI tasks together, with structured outputs, team collaboration, and the kind of orchestration layer that makes multi-step workflows manageable without writing custom code. That experience shaped everything in this book.
For enterprise teams (100+ people), I’d strongly recommend a workflow-building platform like Copy.ai or a similar solution. The bigger the team, the harder it gets to maintain process consistency. Individual contributors will build their own tools, use their own prompts, encode their own understanding of the brand. A dedicated workflow platform forces the process audit from Chapter 13, gives you a shared environment where workflows are visible and editable by the team, and keeps everything organized in one place instead of scattered across personal Claude accounts and homegrown scripts.
For skeleton crews, though, enterprise workflow platforms aren’t always priced for a team of one or two people with a startup budget. The stack below is what works at that scale. It’s more hands-on than a dedicated platform, and some of it requires technical comfort (or a willingness to learn). But it gets the job done at a fraction of the cost.
The honest answer is that the best tool is the one that matches your team size, your budget, and your technical ability. The workflows matter more than what you build them in.
1. AI Backbone
What I use: Claude API (Anthropic) for content generation, transcript processing, outreach drafting, and multi-step workflow processing. GPT-4 for occasional tasks where I need a different perspective or where the training data matters. Gemini for Google Workspace integrations where native access is an advantage.
Why Claude: The output quality for B2B writing is noticeably better. Brand voice matching, nuanced tone control, and long-form coherence are where Claude pulls ahead. For content workflows where the output goes to a prospect or gets published, that quality gap matters.
What I’d change: Nothing on the primary choice. I’d standardize on Claude sooner and stop splitting workflows across models. The overhead of managing prompts across different models creates inconsistency that isn’t worth the marginal benefit of model diversity.
Budget: $20 to $200/month depending on volume (API usage-based).
2. Workflow Orchestration
What I use: FastAPI (custom Python application) for production-critical workflows. This handles the content engine, the repurposing pipeline, and the account intelligence workflows.
Why custom: Full control over prompts, quality checks, error handling, and the sequence of steps. When a workflow runs five times per day and the output goes to prospects, I need to know exactly what’s happening at each step and be able to adjust it without waiting for a vendor’s feature update.
What I’d change: For teams without a developer, I’d start with Make (formerly Integromat) or n8n for workflow orchestration. Zapier works for simple two-step automations but gets expensive and rigid for multi-step workflows. Make gives you visual workflow building with enough flexibility to handle most of the systems in this book. Move to custom only when you hit a limitation that matters.
Budget: Free (FastAPI is open source; hosting on a small server runs $5 to $20/month). Make starts at $9/month.
3. Structured Data and the Brand Brain
What I use: Supabase. The content library, the Brand Brain, the customer language database, the account intelligence records, and the tagged insights all live here.
Why Supabase: It’s a Postgres database with a clean API layer and a generous free tier. SQL access means I can run complex queries when I need to analyze coverage gaps or pull cross-functional data (like which customer language from sales calls appeared in content that influenced closed deals). The API layer means every workflow can read from and write to the same database.
What I’d change: For a non-technical operator who needs something simpler, Airtable or a well-structured Notion database works for the first six months. The limitation is query speed and API flexibility at scale. If your library passes 500 assets and you have five or more workflows querying it, you’ll want a real database. Start simple, migrate when you hit friction.
Budget: Free tier covers most solo operators. Paid starts at $25/month.
4. CMS and Publishing
What I use: Webflow for the content engine’s publishing layer.
Why Webflow: Full control over the SEO technical layer (schema markup, page speed, structured data) without needing an engineering team. The CMS API lets the content workflow push directly to draft status, where I review and publish. The design flexibility matters for AEO because structured content, clean heading hierarchy, and FAQ sections need to be implementable without filing a ticket with engineering.
What I’d change: WordPress with a good hosting provider (like WP Engine or Cloudways) is a perfectly fine alternative, especially if your team already knows it. The SEO plugin ecosystem is more mature. I chose Webflow because I was building from scratch and wanted design control. If I were inheriting a WordPress site with good technical SEO, I’d keep it.
Budget: $29 to $49/month for a CMS plan.
5. CRM and Lead Management
What I use: Salesforce and HubSpot, depending on the property.
Why both: Different properties in my portfolio use different CRMs. My strong opinion isn’t about which CRM. It’s that your CRM must be connected to your content library and enrichment workflows. If lead data, engagement history, and content interactions don’t flow between the CRM and the rest of the system, the inbound and outbound workflows can’t function.
What I’d change: For a startup or skeleton crew starting fresh, HubSpot’s free tier is hard to beat. It covers CRM, basic email, and forms. Salesforce makes sense when you need complex routing, custom objects, or enterprise-grade reporting. Don’t buy Salesforce for a five-person company. Don’t try to run enterprise ABM on HubSpot free.
Budget: HubSpot free to $50/month for starter. Salesforce $25 to $150/user/month.
6. Enrichment and Account Intelligence
What I use: Apollo for contact data and basic firmographics. LinkedIn Sales Navigator for relationship intelligence and hiring signals. Custom scrapers for specific signals (job postings from career pages, recent press releases, tech stack data from tools like BuiltWith or Wappalyzer).
Why multiple sources: No single enrichment provider gives you a complete picture. Apollo has good company data but misses hiring signals. LinkedIn has relationship context but limited firmographic depth. Custom scrapers catch the signals that structured providers don’t index. The account intelligence layer in Chapter 6 needs at least two to three sources to produce briefs worth acting on.
What I’d change: I’d evaluate Clay more seriously for enrichment orchestration. It acts as a connective layer across multiple data sources and can run AI-enrichment steps on top of the raw data. For a solo operator, it could replace the custom scraper work I do manually.
Budget: Apollo $49 to $99/month. LinkedIn Sales Navigator $80 to $100/month. Clay $149+/month.
7. Content Distribution
What I use: Native platform tools. LinkedIn publishing directly. Email through the CRM’s built-in email. Newsletter via ConvertKit. Podcast hosting via a standard host.
Why native: I haven’t found a distribution aggregation tool that’s worth the cost for a skeleton crew. The scheduling and cross-posting tools add a layer of complexity without meaningfully improving reach. The workflow from Chapter 8 handles the repurposing and scheduling logic. The actual publishing is a five-minute task per platform.
What I’d change: If I were managing distribution for more than two brands simultaneously, I’d look at a scheduling tool for social (Buffer or Typefully for LinkedIn specifically). At one brand, it’s unnecessary overhead.
Budget: ConvertKit free to $29/month. Everything else is free with existing subscriptions.
8. Automation Glue
What I use: Zapier for the connections between systems that don’t have native integrations. Post-call transcript to content library. New content published to sales team Slack notification. Lead score change to rep alert. Form submission to enrichment workflow trigger.
Why Zapier: It works. The ecosystem of integrations is the largest, which means fewer “this app isn’t supported” dead ends. The reliability is high enough for production use.
What I’d change: For complex, multi-step automations, I’d use Make instead of Zapier. Make’s visual workflow builder and conditional logic are more powerful. Zapier is better for simple two-step triggers. If you’re building the systems from this book, you’ll outgrow Zapier’s simple triggers within a few months. Start with Zapier for the first connections, migrate the complex ones to Make or n8n as they mature.
Budget: Zapier $0 to $70/month. Make $9 to $29/month.
§The Stack at a Glance
AI: Claude API ($20 to $100/month). Orchestration: FastAPI or Make ($0 to $29/month). Data: Supabase ($0 to $25/month). CMS: Webflow ($29 to $49/month). CRM: HubSpot ($0 to $50/month). Enrichment: Apollo + LinkedIn ($130 to $200/month). Distribution: Native tools ($0 to $29/month). Glue: Zapier or Make ($0 to $70/month).
Total range: $200 to $550 per month. That’s the infrastructure cost for a system that lets one person operate at the output level of a small department.