Start with the process, not the tool
The question is never "where could we put AI". It's "what is somebody doing by hand forty times a week, and why does it have to be a person". Sometimes the answer changes what gets built entirely.
I find the part that's being done by hand, and remove it.
Most businesses don't have a technology problem. They have forty small repetitions a day that nobody has ever stopped to question.
- Works remotely, wherever your business is
- English and Urdu
- You see it running before you commit
Muhammad Rizwan
I design and build AI agents and automation systems — the kind that answer an enquiry in seconds, ask the right questions back, and put the result where your team will actually find it.
What I spend most of my time on isn't the AI part. It's sitting with a process and working out where it actually breaks. Usually it isn't the obvious place. The enquiry gets answered; it just gets answered four hours late. The details get recorded; they just get recorded three different ways by three different people. Once you know exactly where the loss happens, the build is the easy half.
The systems I build are aimed at real estate agencies, clinics and service businesses — places where a slow reply costs real money, and where the same task is repeated every single day. Those are the situations where automation pays for itself fastest.
And I'd rather show than tell. Every system on this site is recorded end to end, so you can watch exactly what happens before we ever talk about building something for you.
How I approach the work
Four things that shape every system I build. They're the reason some projects go quickly and others get turned down.
The messy case is the real test
Any flow works on the tidy example. I build against the enquiry that arrives half-finished at 11pm, the caller who changes their mind mid-sentence, the invoice photographed at an angle. If it survives those, it's ready.
It runs on your accounts, not mine
Your number, your sheet, your calendar, your inbox. You can open the workflow and see every step it takes. Nothing important sits behind a service that only I can log into.
If it isn't worth automating, I'll say so
Some tasks happen too rarely, or vary too much, to be worth building around. It's cheaper for both of us if I tell you that on the first call instead of three weeks in.
What working together looks like
Automation projects go wrong when it's unclear who owns what. So here's both sides of it, plainly.
What you get
- The working system, connected to your own accounts
- A walkthrough of how it runs and where to change things
- The flow written down, so it doesn't live in anyone's head
- Help when a rule needs changing or a new case turns up
What I need from you
- Access to the accounts the system has to touch
- Someone who knows the current process well enough to describe it
- Real examples — especially the awkward ones
- Agreement on what "finished" means before I start building
What I build with
Nothing exotic. These are stable, well-documented tools, which matters more than novelty when a system has to keep running after I hand it over.
n8n
WhatsApp Cloud API
Voice AI
OpenAI
Google Sheets
Google Calendar
Gmail
Google Drive
Airtable
Webhooks & REST APIs
Google OAuth
Tell me where your week goes
Book a free consultation call. Describe the part of your process that feels like wasted effort, and I'll tell you honestly whether it's worth automating.