AI and automation for the work nobody should be doing by hand.
Every business has a list of jobs that go the same way every time: re-typing an enquiry into a spreadsheet, chasing an invoice, sending the reminder, answering the question you've answered more times than you can count. Software should be doing those. We find them, automate them, and use AI where it saves real time, not where it makes a nice demo.
Find the work first
Automation bought before anyone understands the process just makes a mess faster. So we start by listing what actually happens: every repeated task, who does it, how long it takes, where data gets typed twice, which approvals wait on one person. The owner's week is usually the richest source, and the owner is the bottleneck explains why. Then we rank the list by hours saved and risk, and start at the top.
What usually gets automated
- First replies and follow-ups. Instant, in the customer's channel, handed to a person when it needs judgement.
- Data entry between systems. Forms to CRM, CRM to invoices, orders to spreadsheets, without anyone copying and pasting.
- Reminders and chasing. Appointments, unpaid invoices, missing documents, review requests.
- Reporting. The weekly numbers pulled together and sent, instead of someone building them on a Sunday night.
- Handovers. A sale that creates the job, the task and the message to the team by itself.
Where AI earns its place
AI is good at reading, sorting, drafting and answering within limits. It's bad at judgement, and it will confidently make things up if nobody checks. So we give it the first kind of work and keep people on the second: a WhatsApp concierge that answers common questions in two languages, drafts of replies and summaries for a person to approve, tools that sort incoming documents and enquiries. BRZH in Doha runs a concierge that replies in English and Arabic in under five seconds, and a person steps in when it matters. We don't force AI into anything just so it can be called AI-powered.
Built, then run
Automations break quietly: a form changes, a connection expires, a step starts failing at 2am. We monitor what we build, fix it when it breaks, and keep adding to the list as the business changes. That's why automation is part of the engagement rather than a one-off project. Start with the audit if you want the list of what to automate before anything is built, and see the Work page for what's running for clients now.
Who it suits
Owners doing admin in the evenings. Teams copying data from one system into another. Businesses that want to grow without hiring for every new task. Anyone who's been sold an AI tool and isn't sure it did anything. If the business has no CRM yet, start there: CRM and follow-up comes first, because automation needs somewhere to put things.