Service
AI automation that removes real work from your week
We connect large language models and automation tools to the systems you already use, so repetitive work like answering enquiries, preparing quotations or moving data between apps happens on its own.
Where AI automation pays off first
The best first project is usually a task your team repeats every day and can describe step by step. We start there, measure how long it takes now, and automate it.
- Website and WhatsApp assistants that answer product and pricing questions from your own data
- Quotation and enquiry handling that drafts replies for a person to approve
- Reading invoices, forms and emails into structured data
- Internal search over policies, manuals and past projects (RAG)
- Workflow automation between CRM, Google Workspace, accounting and messaging tools
Agents, n8n and MCP servers we run ourselves
We don't just wire up chatbots. We run our own AI tooling in production: a self-hosted n8n instance on Postgres for workflow automation, a self-hosted OpenClaw agent, and MCP servers we built that give AI assistants tools such as web scraping and live exchange market data. The same setups can be built for your business, on your servers or ours.
- n8n workflows connecting your CRM, email, spreadsheets, accounting and messaging
- AI agents (including OpenClaw) set up on infrastructure you control
- Custom MCP servers so Claude, ChatGPT and other assistants can safely use your internal data and tools
- Self-hosted deployments with single sign-on, so workflow data stays on your infrastructure (apart from anything you choose to send to external AI model APIs)
How we build it
We work with models from OpenAI, Anthropic and Google and pick per task on accuracy and cost. Answers are grounded in your own documents, and anything that sends money, signs off a quote or contacts a customer has a human approval step unless you decide otherwise.
Automations run on infrastructure you control, with logs so you can see what each step did.
Pricing
An AI or WhatsApp assistant counts as one main feature on the Custom Build plan, which starts from RM10,000 in Malaysia; international projects are quoted in US dollars. We scope the work and confirm the price before starting. Model usage costs are billed at cost or on your own provider account.
Frequently asked questions
Can you add an AI assistant to my existing website?
Yes. We can add an assistant that answers from your product, service and pricing information without rebuilding the site, as long as we can add a script or component to it.
Is my company data used to train AI models?
We use business API tiers where the providers state that API data is not used for training by default, and we keep your documents in storage you own.
Do you set up n8n or OpenClaw?
Yes. We run both ourselves. We can install and maintain self-hosted n8n or an OpenClaw agent on your server or ours, build the workflows, and connect them to your existing tools.
Do you build WhatsApp chatbots?
Yes. We build WhatsApp assistants on the WhatsApp Business Platform that answer common questions and hand the conversation to a person when needed.