The inbox that answers itself.
Two founders and an agency owner, in the UK and the US, each drowning in email.
- The problem
- Three hundred emails a day between them. The urgent ones hid among the newsletters. Replies took a day because the reading took the day.
- What we built
- A triage system that reads every email as it lands, checks the calendar for clashes, loads the thread for context, drafts a reply in the owner's own tone, labels it, and sends a nudge for the ones that need a decision. The owner reads the draft and hits send.
- What changed
- Under fifty seconds from arrival to a drafted reply. Urgent ones are flagged and nudged the moment they land.
- 300+emails a day handled
- <50sto classify and draft
The lead engine that wakes up at seven.
A B2B founder who needed a pipeline and didn't have a sales team.
- The problem
- Prospecting by hand meant an afternoon lost for twenty emails, most of them to the wrong people, with no follow-up because there was never time.
- What we built
- A daily run that picks a city, pulls matching businesses from Google Maps, checks each one and looks up who runs it, writes a personalised email, sends it through the founder's own mailbox, and follows up for five days unless someone replies. A reply stops everything for that person.
- What changed
- A hundred checked, personalised emails every morning, with nobody typing them. The founder's job became answering the replies.
- 100emails a day, each one written for its reader
- 1manual step: reading the replies
The marketing engine that runs on ten minutes a week.
A content-heavy brand that needed a marketing team's output without a marketing team.
- The problem
- Six channels, each wanting its own format and its own rhythm. Keeping them all fed was a full-time job for three people, so it was nobody's job and most channels went quiet.
- What we built
- One brief on a Tuesday morning goes to a router that hands it to seven specialised writers, one per format. Drafts are checked by a person, then scheduled across all six channels for the fortnight, with reach fed back into the next brief.
- What changed
- Forty-three posts a week across six channels from ten minutes of input. Every channel stays fed, in its own format, without anyone chasing it.
- 43posts a week, six channels
- 10 minof human input a week
The hiring engine that reads every CV properly.
A founder hiring at speed, with more applications than hours.
- The problem
- Every role brought a pile of CVs and a keyword filter that threw out good people for using the wrong words, then kept the wrong people for using the right ones.
- What we built
- A system that watches the applications folder, reads each CV against the actual job description, scores it with written reasons rather than keyword matches, and surfaces only the top few to the hirer. On a yes, it creates the accounts, the welcome note and the first-week plan.
- What changed
- The hirer reads a shortlist with reasons instead of a pile. Onboarding starts the moment someone says yes.
- Top fewsurfaced, with reasons
- 0keyword filters
The seller who could finally see their stock.
A marketplace seller on Amazon, eBay and Temu. A team of two, run from a phone.
- The problem
- Stock lived in boxes across two rooms and a spreadsheet nobody trusted. Things ran out quietly. Orders to suppliers were guesswork, and every week started with counting.
- What we built
- A stock app built for a phone. What is on hand, where it is, what is on the way, and what to order, in plain English. Weekly sales sheets per marketplace. A count mode that logs every check. Push alerts when something is running low or an order is late.
- What changed
- Nothing runs out silently any more. Reordering is one tap from a suggestion, not a Sunday afternoon.
- 4screens, no manual needed
- 1 tapfrom suggestion to supplier order
The founder who posts thirteen times a week.
Our own account, so the numbers are ours to show.
- The problem
- Posting consistently on LinkedIn takes an hour a day that a founder doesn't have, so it doesn't happen, so nothing grows.
- What we built
- A content engine with a memory of every post it has written, a bank of real material to draw from, an image layer, and a feedback loop from reach. It drafts, a person approves, it publishes on schedule to a personal and a company page.
- What changed
- Thirteen posts a week from ten minutes of input. 34,400 impressions in seven days on an account with 700 followers.
- 13posts a week
- 34.4kimpressions in seven days, 700 followers
The front desk that stopped losing evenings.
An appointment-based business in Essex. One owner, a full diary, and a phone that rang during every appointment.
- The problem
- Enquiries came in at night on Facebook and by text. Bookings lived on paper. Anyone who called while the owner was mid-appointment got voicemail, and most of them didn't call back.
- What we built
- A booking website with a front desk that answers the questions people actually ask, takes bookings against a live calendar, sends confirmations, and hands anything unusual to the owner with the details already written down. An admin diary behind it.
- What changed
- Bookings now arrive while the shop is shut. The phone still rings, just not in the middle of a job.
- 24/7answering and booking
- 0evening enquiries left hanging