Parallax Advisory — Inferno Sales Pipeline

Why we build our own

The email-to-action pipeline Eliot is designing is smart. The question isn't whether it works — it's whether Trello should be in the middle of it.

The pipeline you're building

Eliot's design is solid. Emails come in, get parsed, get threaded (three-layer matching — brilliant), get categorized by Claude, and land on a board where salespeople act on them. That flow is the real value.

The board is just where things land. It's the simplest part of the whole system. The intelligence is in the email parsing, the conversation threading, the AI categorization. Trello contributes none of that. It's a place to display cards.

Where Trello breaks the flow

With Trello
Outlook → Parser → Claude AI → ⇌ API → Trello
Every interaction goes through Trello's API. Their rate limits, their data model, their rules.
Without Trello
Outlook → Parser → Claude AI → Our Board
One continuous pipeline. No external dependency. Everything Inferno owns.
Already Built
Sara built a fully functioning project management system with kanban boards, task tracking, and a Q&A messaging system — in one day.
The board isn't the hard part. Your email intelligence layer is the hard part, and you've already designed that.

What "own it" actually means

No per-seat costs
A custom board costs the same whether it's 4 users or 40. Trello's free tier works today — paid tier is inevitable.
No vendor dependency
Trello changed their free tier in 2023. When your workflow depends on a vendor's product decisions, your roadmap is partly theirs.
Data stays in-house
Customer emails, conversation history, pipeline data — all on Inferno's infrastructure. No third-party data processing.
Native integrations
Link a sales card to a quote, a cost estimate, a production order. That's a database join — not an API workaround.
Maintenance is a wash
A Trello integration still needs upkeep — API changes, deprecations, rate limits. You're maintaining code either way.

The real cost comparison

With Trello
○ Free tier now, paid tier later
○ API integration code to maintain
○ Trello's data model constrains design
○ "Does Trello support this?"
○ Customer data on Atlassian's servers
○ Dependent on vendor roadmap
Built in-house
⦿ Fixed cost (~$12/mo VPS)
⦿ Direct database writes
⦿ Data model matches Inferno's reality
⦿ "Does Inferno need this?"
⦿ All data on Inferno's infrastructure
⦿ Roadmap driven by Inferno's needs

Your email processing design is genuinely impressive work. Three-tier conversation matching, AI categorization, attachment handling — that's the product. Don't hand the last mile to a vendor.

Finish what you started.