AI training for your team, and support that keeps you current.

Most AI projects are sold as a build. Someone scopes the work, delivers a system and sends an invoice. Six months later the model it runs on has been superseded, two of the people who were trained have left, and nobody is sure whether the thing is still doing what it should.

We work differently. We help businesses build AI into how they work, and then we stay to keep it current. This page explains why upkeep matters, what to look for in any supplier’s ongoing support, and what we include for our clients. If you are weighing up a first project, start with AI consulting and implementation.

Why an AI system needs upkeep

A spreadsheet macro from 2015 still runs today. AI systems are not like that. Four things change underneath them.

Models and tools change

Model providers release new versions and retire old ones. A system written around one model may need its prompts retested when that model is replaced, because the new one can read the same instruction differently. The automation platforms and business apps around it update too, and a connection that worked last quarter can need attention after a change on either side.

Prices change

Usage pricing moves in both directions. A newer model is often cheaper for the same job, which means a system that was tuned for cost a year ago may now be paying more than it needs to. Nobody notices unless someone is looking.

New capabilities arrive

Features that were not practical at build time become ordinary. Better handling of documents, longer context, cheaper ways to check an answer. A system that was scoped carefully around last year’s limits can often do more now, if someone revisits the scope.

Staff change

The person who understood the system leaves. The new hire inherits a login and a vague explanation. Within a year the team is working around the system instead of with it. Training that happened once at handover does not survive turnover.

None of this means the original build was wrong. It means the build was the start, and something has to keep it current.

What good ongoing AI support looks like

Use this checklist with any supplier, including us. If the answers are vague, the support is probably a retainer without a defined job.

• It is in writing. The support agreement says what is covered, what is not, and what the response commitment is.
• Training is live and on your system. Not a recorded course about AI in general. Your team, your workflow, your examples, with time for questions.
• Someone tells you what changed. When a model you rely on is being retired, or a price or feature changes in a way that affects you, you hear about it before it bites.
• Reviews are scheduled, not reactive. There is a set point where someone looks at the whole system and asks what to upgrade, what to add and what to switch off.
• Removal is part of the job. A good review takes things out. Steps that nobody uses cost money and add risk.
• Changes are reviewed before go-live. An upgrade is tested on your real examples against acceptance checks, the same as the original build.
• The system lives in your accounts. If the relationship ends, you still hold the system and its data.
• You know who to call. A named route to the people who built it, and an honest answer about how quickly they respond.

For the data and model questions that sit under this, see how AppliedAgents handles data and AI.

What AppliedAgents includes with ongoing support

For clients on ongoing support, three things are included. There is no separate price for them.

Team training sessions

Live sessions, in person or by video, with the people who use the system. We cover two things: how to run your own system day to day, and how to use AI tools well in ordinary work, such as drafting, summarising and checking. Because the sessions are live rather than recorded, they fit the way your team actually works. Our guide to prompting AI gives a sense of the ground we cover, and anyone can read it.

Regular AI updates

A recurring briefing on what has changed in AI and what it means for your business. Not a newsletter about everything. A short, specific note: this model is being retired and here is what we will do about it; this price has dropped and here is what it saves you; this new capability is worth a look for one of your workflows, and this other one is not.

Ongoing improvement reviews

Scheduled reviews of your systems. In each one we look at three things: which models to upgrade, which new capabilities are worth adding, and what to remove because it is not being used. Changes are reviewed by senior engineers and tested on your real examples before they go live, as with the original build.

How support works day to day

Senior engineers review all work, including the changes made during a review. Every system comes with documentation, and we walk through it with each client rather than sending a file. Systems are built in your own accounts, and the model provider and region are agreed with you per project, so an upgrade does not quietly move your data somewhere you did not choose.

For clients on ongoing support, AppliedAgents aims to respond within one hour, 24/7. What support covers, and for how long, is set out in your support agreement, and you should ask to see it before you sign, as you would with any supplier.

Who this is for, and who it is not for

The training, updates and reviews are part of ongoing support for AppliedAgents clients. We do not sell them separately, and we do not run training for businesses that are not clients. If you are looking for a standalone AI training course for your staff, we are not the right choice.

What we do offer to everyone is the resource library: free guides on AI for Australian small businesses, written in plain language. If you want to put a few ground rules in place for how your team uses AI before you build anything, the AI policy template is a practical place to start.

Common questions

How often are the AI updates?

They are regular and they follow what is actually changing, rather than a fixed calendar. Ask about the schedule in your first conversation.

What happens when we hire someone new?

Training sessions are part of ongoing support, so they do not stop after the first handover. A new team member can be trained on your system the same way the original team was.

What does this cost?

There is no separate price. Training sessions, AI updates and improvement reviews are included with ongoing support. Your proposal separates the initial build, third-party software and usage, and agreed support, so you can see what each part costs before you commit.

If you want to talk through what keeping your business current with AI would look like, start with a free business audit.