Designing, implementing, and operating practical AI systems for real businesses.

This GitHub is a working portfolio of the systems, patterns, and reference implementations used in client work: agentic operating systems, AI command centres, and domain-specific apps that plug into existing stacks rather than replacing them.

The emphasis is on shippable systems – code that can be adapted to a client environment, not just one-off demos.

• AI implementation and consulting
  Scoping, designing, and shipping AI-powered workflows for operations, support, content, and internal tools.

• Agentic systems and automation
  Orchestrating multi-step, multi-agent flows that combine LLMs, structured tools, and human-in-the-loop approvals.

• Secure, observable, and scalable delivery
  Building with sensible defaults for reliability, logging, metrics, and access control so teams can run these systems with confidence.

Regular tools of the trade include Anthropic, OpenAI, modern web stacks, and integrations into email, chat, CRMs, and data stores.

These are reference implementations, not rigid products. They are designed to be adapted to the tools and constraints of a specific organisation.

• Agentic OS — operating system for AI-powered workflows
  Turns your AI stack into a clickable workspace with agents, memory, notifications, and approvals wired into real processes.
  Good fit for teams who have outgrown “chat with a model” and want structured workflows.

• Command Centre — observability and control for AI environments
  A command-centre style UI that surfaces usage, health, posture checks, and queued tasks, with controls to pause, approve, or override behaviour.
  Good fit for teams who want to see what their AI systems are doing and be able to intervene.

• Guidant 365 — workflow OS for contractors and freelancers
  Tracks jobs from first inquiry to completed work, with stages, next actions, and AI-assisted data extraction under human review.
  Good fit for operators who live in their pipeline and need a clear “what is next?” at all times.

You’ll also find smaller utilities, experiments, and internal tools that explore specific ideas in isolation before they’re rolled into client work.

Most of the code here underpins consulting and implementation engagements – tailoring these patterns to a specific business, stack, and risk profile.

If you are interested in:

• an agentic OS for your workflows
• a command centre on top of your existing AI usage
• or help designing and implementing AI workflows for your operations

you can reach out via:

• Website: https://guidant.nz
• LinkedIn: [https://www.linkedin.com/in/pawaridtupmongkol/]
• Email: guide@guidant.nz