Sorts
Sales, support, accounts, suppliers, spam. Each email lands with the right person.
An inbox that sorts itself and drafts the replies
Most shared inboxes hold the same twenty questions, a few urgent problems and some noise. We build inbox automation that sorts the lot, drafts replies from your own approved answers, and leaves the sending to your team.
What it does
Picks out the details
Names, order numbers, dates and the actual question, ready to use.
Drafts the reply
From your approved answers and knowledge base, with links where they help.
Flags what’s urgent
Complaints, deadlines and anything that mentions a cancellation.
Logs it
Every conversation is recorded against the customer in your CRM.
Before and after
Before
- Everyone reads everything
- The same answers are typed again and again
- Urgent emails are buried
- Nothing is logged
After
- An email arrives
- AI sorts it and picks out the details
- AI drafts a reply from your approved answers
- A person reviews, edits if needed and presses send
A person checks - The reply is logged, and corrections improve future drafts
How we keep it accurate
Measured on your own email
We test sorting and drafts on a sample of your recent email, handled under a data processing agreement.
Corrections make it better
The changes your team makes are fed back regularly, and we check against a held-back set of emails so an improvement in one place doesn’t break another.
Nothing sends itself
Drafts wait for a person. If you later want routine replies sent automatically, we agree which ones, in writing, and log every send.
We run this on our own inbox
Our email is sorted and replies are drafted for a person to check. It’s set up to prepare drafts only; sending stays with a person.
You choose where the AI runs
Tier 1
A hosted model, through a UK or EU endpoint
The best quality for the least effort. Before any of your data goes to a provider, we check that its business terms rule out training on it, and we tell you which region does the processing.
Tier 2
Inside your own cloud account
The model runs in your AWS or Azure account, under your access controls and your logs.
Tier 3
A private model on hardware in the UK
Open-weight models on your hardware or ours. Nothing goes to an outside AI provider. Our own research and document reading run this way.
Email often holds personal data. For most clients we recommend tier 2 or 3.
Storage location and processing location aren’t the same thing. We’ll tell you both, in writing.
What it costs
-
Pilot
From £6,000
Up to six weeks on one job, with the test set and pass mark agreed up front
Running cost estimated per 1,000 emails.
Prices exclude VAT where VAT applies.
Questions
Will customers know it’s AI?
A person sends every reply, so it’s your team’s words. If we build a chat assistant that talks to customers directly, it says it’s AI.
Does it work with Microsoft 365 or Google Workspace?
Yes, both.
What about personal data in emails?
We work as your processor under a written agreement, keep the data where you choose, and can remove personal details before anything reaches a hosted model.
Get in touch
Tell us what lands in your inbox.
One paragraph is enough. We’ll reply with honest thoughts on whether we can help, and what it would cost.