Workflow Automation
Automated pipelines that connect your tools and move data without a human copy-pasting between tabs.
- n8n, Make.com and Zapier workflows
- Error handling, retries and failure alerts
- Sheets, Notion, CRMs, email and REST APIs
I build AI agents, voice assistants, automation pipelines and full-stack dashboards that answer leads, qualify them, update your CRM and hand your team the next action — without anyone touching a keyboard.
Anyone can wire a chatbot that answers a question. I build the boring part underneath: the retries, the edge cases, the CRM writes, the thing that still works on a Sunday at 3am.
Automated pipelines that connect your tools and move data without a human copy-pasting between tabs.
Assistants that understand the request, decide what to do, and actually execute it against your systems.
Voice agents that answer the phone, hold a real conversation, check the calendar and book the slot.
The product around the intelligence: dashboards, directories, portals and real-time apps, built so someone else can maintain them.
Brand sites and launch pages, from a single-page site to a 3D product viewer that runs in the browser.
Agents, voice systems, scraping pipelines, platforms and websites. Filter by what you need built — each one shows the problem, the stack, and whether you can go use it today.
An AI business assistant for real estate agencies. It manages leads, scores intent, tracks follow-ups, searches live market data, keeps property inventory current, and delivers a daily business briefing — all through natural conversation.
Real estate teams bleed leads when follow-ups, property matching and CRM updates all depend on someone remembering to do them.
A working AI employee that agencies can be walked through live, in a real browser, on a real dataset.
Researches a company live on the web, asks qualifying questions, scores the lead, and writes a personalised outreach email.
A caller explains what they need, the agent checks Google Calendar, confirms a slot, books it and sends the confirmation.
A fashion brand site built around a real-time 3D garment viewer. Visitors pick a silhouette and turn the piece around in the browser. There's no product photography anywhere on the page.
Brand direction, copy and front-end build for a UK AI automation agency. A dark single-page site with a 3D hero, six service systems and a carousel where visitors find their own industry. Concept to live in under two weeks.
A lead-generation engine that scrapes six platforms in one run: Instagram, YouTube, LinkedIn, X, Threads and Upwork. Everything is normalised and de-duplicated into a single lead database.
An end-to-end sales pipeline for a UPVC and windows company. Call outcomes land in a post-call sheet, and a quotation engine turns them into a priced quote that goes straight to the customer on WhatsApp.
A voice agent for a Pakistani real estate company. It watches a Google Sheet of new leads and places an outbound call the moment a row is added, so nobody has to dial and nothing waits.
Captures leads from Instagram ads, replies instantly on WhatsApp, collects requirements, and routes each lead to the right team member.
Enriches new leads, scores them, researches recent company news, writes the outreach, then fires a different-angle follow-up 48 hours later.
One idea, video or podcast becomes scripts, voiceovers, short-form video, LinkedIn posts, threads, a newsletter and scheduled uploads, with a human review gate before anything ships.
A voice receptionist that runs in the browser. It answers, holds a natural conversation and routes the caller over WebRTC, with no phone line needed.
The voice half of the lead-to-quote system. It qualifies inbound leads over the phone, captures their requirements in conversation and logs the outcome into the sheet that feeds the quotation engine.
A support agent that follows business-specific instructions to resolve routine queries end to end, and hands off to a human as soon as a question falls outside its scope.
Computer vision on live camera feeds that sends a WhatsApp alert the moment it spots vaping. Built for schools and managed facilities.
A platform for discovering and supporting Black-owned businesses across the US. Searchable directory across 20+ categories, business profiles, direct messaging and a paid listing tier.
A real-time messaging app with DMs and groups, replies, reactions, attachments, read receipts, typing indicators and live presence. Everything updates over the socket without a refresh.
A mental-wellness platform with mood tracking, journaling, an AI support chat, real-time emotion detection from webcam and voice, and a local support directory.
An online learning and exam-prep platform with AI tutoring and a multiplayer whiteboard for group study sessions.
A block-editor task manager built on one rule: work flows down the hierarchy and nothing gets lost. Fully self-hosted, with no cloud account and no API keys.
Time-series models for supply-chain inventory planning, built through repeated rounds of data cleaning, feature engineering and model selection.
A brand site built around 3D scenes and scroll-driven animation. The browsing itself carries the brand, more than the copy does.
The goal isn't just a system that works — it's a system you understand well enough to explain to your own team without calling me.
I map your current process and find where automation saves the most hours, not the most impressive demo.
I design the workflow, tools, data flow, edge cases and failure handling — before writing anything.
I build and test against real scenarios and messy real data, not the happy path that always works in a demo.
You get documentation, workflow access and a recorded walkthrough. The system is yours, not rented from me.
Send me the workflow, your tool stack, or just describe the manual thing that eats your week. I'll tell you what can be automated, what shouldn't be, and what the first useful version looks like.