A small agency that builds things that work.
RelayOps is a one-person AI automation studio based in Nairobi, Kenya. We build systems that handle the operational gaps — slow lead follow-up, manual appointment booking, repetitive admin — for agencies and service businesses that can't afford to hire their way out of the problem.
Why RelayOps exists
The problem that led to RelayOps is simple: agencies and service businesses lose deals not because they're bad at what they do, but because they're slow to respond. A lead submits a form at 9pm. Nobody sees it until the next morning. By then, whoever replied first — at 9:01pm — has already had a conversation.
The common advice is to hire someone to manage leads. But that doesn't solve the structural problem — it just adds headcount to a process that's still fundamentally manual and still breaks at 9pm on a Friday. The right answer is a system, not a person.
RelayOps was started to build those systems. Not as flashy AI demos, but as reliable operational infrastructure — the kind that runs in the background without anyone needing to babysit it, and that hands off cleanly to a human when a human is actually needed.
Brian Chebitoch
·Founder & Automation EngineerIT and systems engineering background. Studying Information Technology at Multimedia University of Kenya. Approaches AI automation as an engineering problem, not a sales pitch — which means the systems are designed to be reliable, documented, and possible to maintain without a technical team standing by.
What we're trying to do
Build AI automation systems that are reliable enough to trust, transparent enough to understand, and documented well enough to outlast the person who built them.
That sounds simple. It's harder than it sounds. Most automation projects fail not because the technology doesn't work, but because the system was built for a demo, not for daily operation. RelayOps exists to close that gap — to build things that are still working six months later, that the client understands, and that don't require a technical co-founder to maintain.
How we operate
Honest over impressive
RelayOps is a small, new agency. We don't pretend otherwise. Every page on this site reflects what we've actually built or can demonstrably build — not what sounds good in a pitch.
Reliable over clever
A system that works quietly every day is worth more than a demo that impresses once. We design every workflow to handle failure gracefully — and to be easy to understand months later.
Direct access, always
There's no account manager between you and the person building your system. You talk directly to Brian throughout — during scoping, during the build, and after it's live.
How we think about tools and AI
Opinions developed from building systems that had to work in production, not in a sandbox.
Tools are chosen for the job, not for the portfolio
n8n if you want self-hosted control, Make.com if you want a managed platform, Zapier if your team already knows it. The workflow logic is what matters — the wrapper is secondary.
AI is used where it genuinely helps, not everywhere
GPT-4 for qualifying conversation and intent classification. Structured logic for routing and CRM writes. Using LLMs for things rules handle better creates unpredictability.
Every integration is treated as a potential failure point
Webhooks time out. APIs return unexpected formats. Rate limits hit at 2am. Every build includes error handling for the things that will definitely go wrong, not just the happy path.
Clients should be able to understand what they've paid for
If the documentation requires technical knowledge the client doesn't have, it's not good documentation. Every handover is written for the person who will actually use it.
The delivery process
Every engagement follows the same structure. Steps may compress or expand depending on scope — but none are skipped.
Discovery
30–60 minA single call to understand your current workflow, where time is being lost, and what a working solution actually needs to do. No commitment required.
Workflow Audit
AsyncWe map the current process end-to-end — every manual step, every handoff, every tool involved. This is where the real problems surface, not what was described in the call.
Solution Design
2–5 daysA written spec covering the proposed automation, the tools used, what it handles, what it escalates to a human, and what it doesn't do. You approve this before any build starts.
Build
1–3 weeksThe actual build — workflow logic, integrations, conversation flows, error handling. Progress is shared as it develops, not revealed as a final deliverable.
Testing
3–7 daysThe system is tested against real-world edge cases — duplicate submissions, API failures, unexpected inputs, off-hours behaviour. Issues are fixed before your team sees it.
Deployment
1–2 daysGoes live with a rollback plan. We monitor the first 48–72 hours closely and stay available for anything that surfaces in production.
Documentation & Handover
IncludedEvery build ships with a plain-English guide covering what it does, how to make common changes, what to check if something breaks, and who to contact. Not a diagram — a document your team can actually use.
Ongoing Support
OptionalAvailable for monitoring, tweaks, and expansion. Some clients want a retainer; others prefer to call when they need something. Both are fine.
What RelayOps is right now
Brian handles scoping, building, testing, and support directly.
No past client logos to show — but the builds are documented and demonstrable.
A small number of new projects per month to maintain quality.
Works with clients globally. EAT timezone (UTC+3). Async-friendly.
Why the transparency? Because agencies vetting vendors do their homework, and a site that overstates its history gets caught immediately. RelayOps is small and new — and that's fine. Being small means direct access to the person building your system, a genuine interest in making it work, and no incentive to oversell scope.
Book a discovery call
Tell us what's eating up your team's time. We'll tell you honestly whether automation is the right fix — and what it would involve.
Schedule a callView demo projects
See a fully-documented build — architecture, workflow, tooling, and expected outcomes — before you commit to anything.
View demo projects