Why Parenza

Before our first child was born, I did what I do at work: I wrote things down. Not a diary — a structure. What we'd decided and why. What the NHS guidance actually said versus what the internet said it said. Which questions were feeding questions, which were sleep questions, and which only looked like one and were really the other.

Then the baby arrived, and the structure earned its keep. Every night we logged feeds and sleep. Every hard call — the reflux question, the vitamin D dosage where UK and Russian guidance disagree, when to start the nursery waitlist — went into a decision log with the reasoning attached. Within weeks there were 1,300+ logged entries and 44 documents of family context behind them.

Who I am, and why it shows in the product

I'm Anton Batiaev — VP Product & Engineering, Capital.com, and before that Revolut, Deutsche Bank and a fintech I took from zero to launch as CTO. Fifteen years of building systems where being approximately right is the same as being wrong, plus a master's in computer and information systems security. More at batiaev.com.

Three things from that work ended up inside Parenza.

I work with AI all day. Not as a novelty — as the tool I reason with. Which is why Parenza was never going to be a walled app with a chatbot bolted on. If the assistant you already use can read your family's actual record, that's worth more than another icon on your phone. That's what the MCP connection is for.

I've been through the audits from the inside. SOC 2, ISO 27001, DORA — in businesses where a regulator asks for evidence and "we intend to" is not an answer. That experience is why the trust page reads the way it does: control by control, with honest statuses, and no badge we haven't earned. Data governance for a child's health record deserves at least the seriousness a bank applies to a balance.

I validate before I act. In risk systems you learn quickly that a confident number from an unchecked source is the most dangerous thing on the screen. Parenting advice has exactly that problem, at scale, aimed at exhausted people. So the specialists here name what they're working from, flag where national guidance diverges, and say plainly when a question needs a clinician rather than an answer.

The part that actually worked

The logging wasn't the valuable bit. The valuable bit was pointing an AI assistant at all of it and asking a real question — and getting an answer that knew this child's history instead of an average child's. Then splitting that into 10 specialists with distinct remits, each with its own sources and its own stated limits, and putting a Chief of Staff above them to route questions and reconcile disagreements.

That's not a metaphor for how Parenza works. It is how it works, because it's what has been running on our own family since July 2026 — before any of it became a product.

Why it isn't a tracker

Trackers are built around the log. You put data in and get charts back, and the charts don't help at 3am, because at 3am you don't have a data question — you have a decision. Parenza is built around the decision. The log exists to make the decision better.

Who it's for

Parents who want the reasoning, not just the reassurance — people who are comfortable being told "the guidance says X, your history says Y, here's the trade-off" and making the call themselves. If what you want is a soothing voice telling you it's all fine, there are gentler products than this one.

Where it is now

Pre-launch, with a small design-partner cohort. No billing, no public signup, and no claims about scale we haven't reached. If this is how you'd want to make these decisions, tell us about your family on the early-access form, or email hello@parenza.ai.

Employers are named as professional background only. Parenza is an independent project, not affiliated with or endorsed by any of them.