Talk numbers
Work

Client stories: the question each business couldn't answer, what we built and what changed.

All work
Built

Tools we built, for clients and for ourselves, when nothing we could buy did the job.

Open to anyone

Runs our work

For clients

See all builds
AI

AI moves fast. Here's how we use it in our own work, and where to start in a business your size.

The AI hub
Approach

Every business is different, so we always start with a conversation.

How we work
  1. 15 minA quick callYour goals, your numbers, and whether we're a good fit.
  2. 1 hourDiscoveryWhere you're losing enquiries, and what's already working.
  3. ThenA clear planPaid ads, a custom build or both, as a one-off project or ongoing.

What we connect

About

We're a small London team of marketers, and we use AI to go faster.

About us
  • Cam BowenCam BowenCo-founder
  • Ben PowerBen PowerCo-founder
  • ElectraElectra
  • MahirMahir
  • NatashaNatasha
  • RahabaRahaba
  • ZobaerZobaer
AI updates

What we've changed lately, newest first.

Every change to our AI setup and the way we work, with the date. Each one says what we changed and why.

Back to the AI hub
The list

Everything we've changed

RSS

  1. Rule

    Every repo now gets a secret check before each commit

    Before each commit in our working copies, a check stops anything that shouldn't be saved, such as keys. We first added it on 25 Sept and rolled it out to every repo on 8 Oct. A person can still override it.

    Read more

    Our AI tools write a lot of files, so we added a routine safety check.

    What changed

    • A hook scans what's about to be committed.
    • If it finds text that should stay private, the commit stops until it's fixed.
    • It also strips keys out of the saved AI chat transcripts we keep.
    • It's opt-in for each working copy, and it's switched on when a Claude Code session starts.
    • A person can still override it on purpose, so it's a safety net, not a guarantee.

    Keys and passwords stay in environment files and never get committed.

    CursorClaude Code

  2. Rule

    A hook now stops Claude Code sending email or chat for us

    Our rules have said since August that AI tools can draft an email or a chat message but don't send it unless a person asks. On 25 Sept we added a hook that enforces it in Claude Code, for sending, inviting, sharing and publishing.

    CursorClaude Code

  3. Setup

    Cursor and Claude Code now follow one set of rules

    We now generate Claude Code's rules, tool connections and agents from the Cursor ones, so both tools follow the same house rules. A check tells us if the two drift apart.

    Read more

    We used to keep two sets of rules, and they kept drifting apart.

    What changed

    • The Cursor rules are the only ones we edit.
    • A small script makes the Claude Code copy from them.
    • A check fails if the copy stops matching.

    So British English, no em dashes and the no-send rules work the same way whichever tool someone opens.

    CursorClaude Code

Get updates

Get each AI update by email.

We'll email you now and then, when we change a workflow, a tool or a verdict. You can unsubscribe any time.

By subscribing you agree to our Privacy Policy.

Next step

Bring your numbers.

We'll tell you if the spend makes sense.

  1. A 15-minute callYour goals, your challenges, and whether we're the right fit.
  2. A 1-hour discovery callWe dig into your numbers. Worth a hundred emails.
  3. A concrete planAds, a custom build or both, as a one-off project or ongoing.
Talk numbers
15 minutes · No pitch deck · A straight answer