What we automated in twelve months

01

Building the main CRM board meant copying records across by hand, one field at a time. A desktop app now reads profiles from a public industry database and writes ten fields straight onto the board, including valuation, key dates, location, representation, contract and supplier.

02

7,560 rows of sales figures converted from pounds to dollars, re-runnable at a new rate without doubling anything up. Every change logged. The original pound figures were never touched.

03

Top tier, Active and Inactive were sorted by hand, from spending split across two boards in two currencies. It now adds up both, converts the pounds so the totals compare, and moves each client itself. You see who is about to move, and why, before anything changes.

04

Someone on the team exported clients from Monday as a CSV, tidied it and uploaded it to Wati. It worked, but somebody had to remember, and between exports the list was out of date. All 3,000+ clients now feed through automatically, linked family members and partners included, with phone numbers corrected on the way so messages stop silently failing.

05

Reps label a chat in the WhatsApp Business app they already use. It becomes an item in that manager's group on the enquiries board, checked first so nothing is logged twice. No new habit, no writing anything up.

06

Deal notes lived in reps' heads and cards sat in the wrong stage for weeks. The system now reads each WhatsApp conversation, writes the notes onto the card, pulls out the value being discussed and moves the card to the right stage. It will not invent a figure the chat never mentioned.

07

A decade of relationships sat unsearchable in Outlook. We read 190,000 messages across 21 mailboxes back to 2015, built one record per person, pulled phone numbers and job titles from signatures, and tagged each company by industry. Pick an industry, see every brand, see who you know at each one.

Industry tagging matches the client's own team 79 percent of the time, tested against 532 accounts they had already labelled. Contacts who never replied leave no signature, so six in ten records carry a name and three in ten a phone number.

08

That database would have gone stale immediately. Now every email arriving in the 21 inboxes is checked against the board, and anyone new is added from their signature. Internal mail, newsletters and automated senders are filtered out, and it runs unattended.