Consulting
I’m Brent Burgess, a data architect in Benicia, California. I help small teams whose numbers have outgrown the spreadsheet: a data platform you can trust, and AI that works on it without being handed everything.
Three ways to work together, each with a fixed scope and a fixed price agreed before it starts. Each one below describes a situation, not a past client, and links to the same kind of system running on this site.
To start, email brent@brentburgess.com with the tools you use now and the report or task that takes too long. I reply with whether the audit fits and what it would cover.
Architecture and automation audit
- Where you are
- Your numbers live in a few spreadsheets, an accounting system and a sales tool, and the monthly report is a day of copying between them. Two people disagree about a figure and neither can say which is right.
- What I would build
- A written plan: a map of your sources and where each number should come from, the tables and pipelines to build, in what order, and which repetitive task to automate first.
- What you get
- The plan, a walk-through of it, and a first estimate for each piece, so you can build it with me or with anyone else.
- Typical length
- One to two weeks.
See it in practice: the pipeline behind this site’s own data.
What the written plan looks like: a sample audit plan, written about that pipeline.
Build sprint
- Where you are
- You have the plan, or one report that already takes too long: an export, a clean-up and a paste, every week.
- What I would build
- One pipeline or one automated workflow from the plan, with tests that fail loudly, monitoring, and handover notes your team can maintain.
- What you get
- The running system in your own accounts, the notes, and a handover session.
- Typical length
- Two to four weeks.
See it in practice: how a Sleep Anchor track ships.
AI access to your data
- Where you are
- Your team asks the same questions of the same data, and the answer depends on who has the spreadsheet open. You want an AI assistant on it without handing it everything.
- What I would build
- A read-only connection between your data and the AI tools your team already uses, with what it may see written down, every answer traceable to its source, and checks on what comes back.
- What you get
- A connection your team can ask questions through, a page that says what it can and cannot see, and the checks that run on its answers.
See it in practice: this site’s own read-only MCP server.
Your data, and the AI’s output
Where the data lives: in your own accounts. I work inside them with the least access that does the job, and nothing is copied out to mine.
Who can reach it: the people and tools you name, each with its own credentials, and read-only ones for anything that only needs to read. This site works the same way: its analytics pipeline runs on a token that can only read, and its deploy token can change one Cloudflare Worker and nothing else.
What checks the AI’s output: tests that fail loudly before a number is published, and a person who signs off on anything that cannot be undone, which is how every Sleep Anchor track ships.
How it works
- Fixed price, quoted and agreed before work starts.
- Most engagements start with the audit.
- Larger or custom work: email me with what you have in mind.
What to put in your first email
- The tools and data sources you use now.
- The report or task that takes too long.
- What a good result would look like.
Those three, so the first reply can say whether the audit fits and what it would cover.
To talk about an engagement, email brent@brentburgess.com. BurgessData.com points here.