AI systems, built by a senior engineer. Not slide decks.
I'm Farzin. Xerad is an engineering studio I run out of Vancouver, building AI agents, automations, and data systems for teams that want to move faster — without taking on the operational burden of running them. We build, host, and operate what we ship. You focus on the value it produces.
Build-as-promised, or your money back.
Every proposal spells out exactly what your system will do, in writing. If the finished build doesn't meet that specification, I have 30 days to correct it — and if it still doesn't, you get a full refund of what you paid. Fixed scope, fixed price, agreed before you start.
Built by a senior engineer with 10 years of production experience.
Most projects ship in 4–6 weeks, fixed price.
Fully managed — hosted, monitored, and operated by Xerad.
Fixed scope, fixed price, fixed timeline — in writing before you commit.
Farzin — Founder · Senior Engineer
- ALM (Master's) in Data Science, Harvard (candidate)
- Ten years building production systems
- Based in Vancouver, Canada
A senior engineer leads every engagement at Xerad — not an account manager. That engineer is me.
Operations don't break overnight.
They erode quietly.
Manual work eats your team
Your best people spend hours on tasks an AI agent could finish in seconds — copying data between tools, answering the same questions, chasing approvals that should be automatic.
Your data sits silent
You're collecting customer signals, ops events, financials — but it lives in spreadsheets and SaaS dashboards no one reads. The insight is there. The system to surface it isn't.
Off-the-shelf doesn't fit
Generic SaaS forces your business to bend around its workflows. The opposite is now possible: build the system around how your team actually operates.
Example workflows I'd build
A few representative engagements — the kind of work Xerad is built to do.
WhatsApp intake & qualification, in Farsi or English
An immigration consultant gets 30+ WhatsApp messages a week from prospective clients — half don't qualify, the rest get answered in fragments between case work. A bilingual agent handles intake, asks qualifying questions, captures details, and hands off ready prospects with full conversation context. Same pattern works for clinics, brokers, and legal practices — anywhere DMs are the first touchpoint.
Document extraction for paperwork-heavy practices
Tax season at an accountant's office means hundreds of T-slips, receipts, and supporting documents arriving by email and WhatsApp. A pipeline reads the incoming documents, extracts structured fields into the client file, flags what's missing, and queues clean records for human review. Built around the document types you actually handle — not a generic OCR tool you'd have to configure yourself.
Sarafi reconciliation: Interac in, wires out
A money services operator runs dozens of Interac e-transfers in and wire requests out every day, then spends Sunday night matching them against the ledger by hand. An automated reconciliation flow pairs transfers to client orders, flags discrepancies for review, and produces the audit trail without the manual cross-referencing.
Bilingual booking & reminders for appointment-driven businesses
A med spa or dental clinic loses revenue to no-shows and to English-only confirmation flows that older clients ignore. A WhatsApp-first booking agent takes new appointments, sends reminders in the client's language, and handles reschedules — without the front desk chasing anyone manually.
Audit → Proposal → Build → Deliver
Audit
A free working session to map your workflows, surface the bottlenecks, and identify the highest-leverage place to start. You leave with a plan — whether or not we build it.
Proposal
Within three business days, you get a written proposal: fixed scope, fixed price, fixed timeline. No hourly billing, no scope that quietly expands. If there isn't a fit, I tell you why.
Build
We design and ship the system against your real data, in your real environment. Most projects ship in four to six weeks, hosted on our infrastructure and wired into your stack.
Deliver
You sign off only when the system does what the proposal said it would. If it doesn't, I fix it — or you get your money back. Documentation and a clean path to take it in-house are included.
What you actually get when you hire Xerad
"We don't know where to start"
You don't need to. We audit your operations, find the highest-impact places to apply AI, and hand you a roadmap with realistic timelines and dollar figures. No buzzwords, no waste.
"Will it work with our stack?"
We integrate with what you already run — CRMs, billing, communication tools, databases, internal apps. If your team uses it, we can plug into it.
"What if AI changes again?"
It will. We design for change — we host the system and handle model and provider swaps ourselves. The system keeps running while AI churns underneath.
NestM8 — AI buyer-qualification agent for real estate
A working AI conversational agent in production for the Vancouver real estate market — built end-to-end as an internal lab project.
NestM8 is where we de-risk patterns before bringing them into client engagements.
See it work →
Structured access to senior engineering
before the case studies are written.
I'm taking 5 companies into Xerad's founding cohort. In exchange for reduced pricing on a pilot engagement, you get the founder on every call, direct input into how the studio works, and your project becomes a featured case study when Xerad publishes its 2026 portfolio.
This isn't a discount because I'm new. It's structured access to senior engineering at a rate that won't exist once these case studies are written.
Projects are fixed-price, in writing, before you pay. The audit is free.
What people ask before we start
What does this cost?
Most projects land between $3,000 and $8,000, fixed price. You see the exact number in a written proposal before you commit to anything — no hourly billing, no scope that quietly expands. The audit that produces the proposal is free.
How do we start?
Book a free audit. If there's a fit, you get a written proposal — fixed scope, fixed price, fixed timeline — within three business days. If there isn't a fit, I'll tell you why.
What if the audit shows you can't help me?
Then I tell you that, and we don't go further. The audit is genuinely diagnostic, not a sales formality. “This isn't worth building right now” is a real possible outcome, and you've lost nothing but the call.
What if the system doesn't work?
Every proposal spells out exactly what your system will do, in plain, specific terms you can check against. You review the finished system against that definition before the project closes — so you're never paying for something unfinished. If it doesn't meet the specification, I have 30 days to correct it. If it still doesn't, you get a full refund of what you paid.
A refund unwinds the project — the system is deactivated, since you're no longer paying for it. That's the only condition: a refund and a running system aren't both on the table.
The one thing this doesn't cover is business results. How many customers you get depends on your traffic and your follow-up, which aren't mine to control. I'm direct about that line up front, because a guarantee only means something if it's specific.
I've worked with developers or agencies before and it didn't go well.
Fair — that's common. Two structural differences here: scope, price, and timeline are fixed in writing before you pay, so the project can't quietly expand. And the work is project-based, not a monthly retainer you have to remember to cancel. You're buying a defined thing, not an open-ended relationship.
Do I own what you build?
Yes — your system, your configuration, and all your data are yours. You're not renting access and you're not locked in. The one thing I keep is the underlying framework I build on, the same way any developer reuses their own tools across projects. What's specific to your business is fully yours.
Do I need to understand AI to work with you?
No. The audit and proposal are in plain language — what the system does, what it costs, what changes for your customers. The technical decisions are mine to carry.
Why should I trust a small company with this?
A reasonable concern, and it cuts both ways. The honest version: you get the people who build it on every call, and nothing gets lost in handoffs — but there's no large team behind it. What I can show you is NestM8, a production AI system built and run end to end. You can see a real system working before you decide.
Can you build this for Farsi-speaking customers?
Yes. Building bilingual systems for the Farsi-speaking community is a core part of what Xerad does.
Have a workflow worth fixing?
Tell me what's slowing your team down. I'll come back with a short read on whether it's a good fit and what a first engagement could look like.
Start a conversation →