getting started

+61 2 5500 9500

Mnesis Logo

How It Works

How Does a Custom AI Assistant Work?

I build a private AI assistant trained on your business — your documents, your processes, your terminology. From a discovery call to a fully deployed system, here's exactly what happens at each step.

Built for your business Cloud or private deployment Fully managed for Australian businesses
Get Started Use Cases
Discover
Audit
Build
Integrate
Deploy
Train
Refine

The Process

From Discovery Call to Deployed AI — Step by Step

Every custom AI assistant is built from scratch for your business. No templates. No generic setups. Here's the exact process from start to finish.

1
2
3
4
5
6
7
Step 1

Discovery Session

It starts with a conversation. I learn how your business works — your workflows, your pain points, the tasks that eat up your time. I ask about the tools you use, how your team communicates, and where information lives. This call shapes every decision that follows.

Discovery Call In Progress
Workflows mapped 5 identified
Pain points 3 critical
Tools in use 4 systems
Step 2

Business Audit

After discovery, I audit your business for the highest-impact AI opportunities. Not everything needs automation — I identify the tasks where AI will save the most time, reduce the most errors, and deliver the clearest return. You get a prioritised list before anything is built.

Example: 6 hrs/week at $80/hr = $24,960/year returned to your business.

Impact Assessment Complete
  • Client report drafting — 6 hrs/week saved
  • Research and compliance checks — 4 hrs/week
  • Email and client comms — 3 hrs/week
  • Quote and proposal generation — 2 hrs/week
Step 3

Workspace Build

This is where the AI becomes yours. I build a structured workspace with custom instructions, organised knowledge bases, and business-specific prompts. Your documents, templates, terminology, and processes are loaded so the AI understands your business from day one — no retraining needed every conversation.

AI Workspace Building
  • Custom instructions configured
  • Knowledge base loaded (14 documents)
  • Business terminology mapped
  • Prompt templates built (8 tasks)
Step 4

Tool Integration

The AI connects to the tools you already use — calendars, email, CRMs, accounting software, project management. It doesn't replace your systems. It works inside them, reading and writing data where you need it. No new logins, no new dashboards.

Connected Tools
Calendar
Email
CRM
Accounting
File Storage
Project Mgmt
All tools connected
Step 5

Cloud or Private Deployment

Two paths, depending on your needs. Cloud-based uses Australian data centres with enterprise-grade encryption — fastest to deploy, lowest barrier. Private deployment installs AI models on your own hardware for full Australian Privacy Act compliance — zero data leaves your network. File-based setups give you complete control with no external connections, ideal for air-gapped environments. I recommend the best fit during discovery.

Deployment Deploying
Environment Cloud-based
Workspace configured Yes
Integrations active 6 of 6
Knowledge base 14 docs loaded
Step 6

Training and Handover

You get a live walkthrough of every capability I've built. I provide written SOPs, video guides, and a verification checklist so you can confirm everything works the way you need it. You'll be confident using the system before I hand it over — no guesswork.

Training Package Delivered
  • Live walkthrough session completed
  • Written SOPs for every task
  • Video guides recorded
  • Verification checklist signed off
Step 7

Ongoing Refinement

Your AI assistant gets better over time. On a monthly retainer, I update the knowledge base, add new capabilities, refine prompts based on how you use it, and keep everything running smoothly. As your business grows, your AI grows with it.

Start with one workflow, prove the ROI, then expand. I recommend testing with the highest-impact task first, measuring the results, and adding more capabilities once the value is clear. No risk, no guesswork.

Monthly Report
Tasks automated 127 this month
Hours saved 15+ hrs/week
New skills added 3 this month
Knowledge base 22 documents

Deployment Options

Two Ways to Deploy Your AI Assistant

Every business has different requirements around speed, data control, and budget. I recommend the best path during discovery — and can migrate between them later if your needs change.

Cloud-Based

Cloud-Based Setup

I configure commercial AI platforms with your business context, custom prompts, and integrated workflows. You access your AI through tools you already know — no infrastructure to manage.

  • Fastest to deploy — days, not weeks
  • No hardware required
  • Always updated with latest AI models
  • Lower upfront cost
Best For
Most businesses. Fast deployment, low barrier, immediate results.
Private

Private / Local Deployment

AI models installed and running entirely on your own hardware or private cloud. Zero data leaves your network. Full control over the model, the data, and the infrastructure.

  • Complete data sovereignty
  • No third-party data processing
  • Works offline — no internet required
  • Full compliance with data regulations
Best For
Legal, financial, and compliance-heavy businesses with strict data requirements.

What You Get

Everything Included in Your AI Assistant Build

Every build is custom, but here's what's standard across all engagements. You get a complete, working system — not a half-finished prototype.

Custom AI Workspace

Structured folders, custom instructions, and organised knowledge bases tailored to how your business operates.

Business Knowledge Base

Your documents, templates, terminology, and processes loaded into the AI so it understands your business from day one.

Tool Integrations

Connected to your calendar, email, CRM, and existing software. The AI works inside your tools, not alongside them.

Custom Prompt Templates

Pre-built prompts for your most common tasks — reports, emails, research, documentation. One click to execute.

Training Materials

Written SOPs, video walkthroughs, and a verification checklist. You'll be confident using the system from day one.

Ongoing Management

Monthly retainer for updates, new skills, knowledge base maintenance, and expanding capabilities as your business grows.

See how businesses use their AI assistant

Common Questions

Common Questions About the Setup Process

Cloud-based setups typically take one to two weeks from discovery to deployment. Private/local installations take longer — usually three to four weeks, depending on the hardware and complexity. I provide a specific timeline in your proposal after the discovery call.
Not at all. I handle all the technical setup. You interact with the AI through a simple interface — type a question or select a task, and it handles the rest. The training session covers everything you need to know, and the written SOPs are there for reference.
Virtually anything text-based. Standard operating procedures, policy documents, contracts, email templates, product catalogues, client communication guidelines, compliance documents, training manuals — if your team references it, the AI can learn from it.
Yes. Cloud-based setups use enterprise-grade platforms with encryption in transit and at rest. If data sovereignty is critical, the private deployment option keeps everything on your own hardware — no data leaves your network. I discuss security requirements during discovery and recommend the appropriate path.
In most cases, yes. Calendar systems, email platforms, CRMs, accounting software, project management tools, and file storage are all common integrations. During discovery, I audit your current tools and confirm what can be connected directly.
You move onto a monthly management retainer. I update the knowledge base, add new skills, refine prompts based on your usage, and handle any technical maintenance. Your AI assistant improves over time, not just at launch.
Yes. All builds are structured to be portable. Your knowledge base, custom instructions, and prompt templates transfer between environments. If your requirements change — say you take on a client with strict data residency rules — I can migrate your setup.
Your AI assistant is grounded in your own business data — it draws from the documents, templates, and processes you provide, not from the open internet. Every output is designed to be human-reviewed before action. The AI assists and accelerates your work, it doesn't operate autonomously. I also build in guardrails during setup to keep responses within the scope of your knowledge base.
A full-time admin costs $55,000+ per year. An in-house AI engineer runs $150,000+ per year. A custom AI assistant is a fraction of both — and it works 24/7 without sick leave, superannuation, or training gaps. You get the same output at a fraction of the overhead, with ongoing management included in your retainer.
Absolutely — and I recommend it. A phased approach lets you prove the ROI on your highest-impact task first, measure the results, and then expand with confidence. Most clients start with one or two workflows and add more once they see the time savings in practice.
Still have questions? Book a free discovery call

Ready to Build Your AI Assistant?

Book a free discovery call and I'll learn how your business works — then map out exactly how a custom AI assistant would save you time every week. Learn more about Mnesis.

Get Started Use Cases

Fully managed service. No lock-in contracts. Cancel anytime.