Data Architect · Pro
An AI data architect for model and pipeline reviews
Data design decisions get made out loud and then lost. The Data Architect persona forces the questions that make a model reviewable six months later: which system is the source of truth, who writes versus who reads, how the entities actually relate, and whether any of it is personal data with a retention rule attached. Each decision comes back as an ADR carrying the context that produced it.
What she does in the room
Your one job is to capture the data architecture decisions made in this meeting, and record them. That is the whole role.
Once oriented, force clarity on system of record, who writes vs reads, entity naming, cardinality, and PII or retention rules. Ask which system owns each concept before diving into attributes.
The questions she actually asks
She opens by working out what the session is for, then presses on whatever is vague or unowned. These are her real prompts, not paraphrases:
- “Which systems or data domain are we discussing today?”
- “What is the group hoping to settle about the data in this session?”
- “Are there constraints already known — governance, security, or compliance — that I should listen for?”
- “Which system is the source of truth for that?”
- “Who writes it, and who reads it?”
- “Is any of that personal or sensitive data?”
- “What is the retention or governance rule?”
What she listens for
Across the meeting she is collecting schema and data-model decisions, architecture decision records (ADRs), data flows between systems, interface contracts, master-data ownership, data-related constraints (retention, governance, security, compliance, PII), decisions, action items, and risks / dependencies / assumptions / constraints. She records only what participants actually said — when something is missing she asks rather than fills it in.
Captured as structured items
- Schema decision
- Architecture decision record
- Data flow
- Interface contract
- Master data ownership
- Data constraint
You walk out with
- Data model & decisions
- Integration notes
- ADR log
Session presets
- Architecture review
Meetings people use this for
- Data model reviews
- Pipeline and integration design
- Data governance sessions
- Architecture review boards
Getting her in the call
Paste a Teams or Zoom join link, pick Data Architect, and she dials in as a named participant — no tenant install and no admin consent. She introduces herself as Guy by default; you can rename her and change her voice per persona in settings.
Included in the free 60-minute trial, then part of Pro at $39/month. General Notetaker and Business Analyst stay free.
Adjacent personas
Different meeting, different seat. These are the personas closest to this one: