Overview
What generative AI is good for, and where Great Minds fits
Generative AI models produce text, images, code and summaries from instructions. In business, the applications that pay back are rarely chat for its own sake. They are drafting first versions of documents and replies, answering staff questions from your policies and knowledge base, summarising cases, calls and long threads, extracting data from documents, and producing channel copy or translations for a person to review.
Organisations use generative AI in three ways: an assistant for their people, such as ChatGPT, Claude, Microsoft 365 Copilot or Gemini; models built into platforms they already run, such as Atlassian Rovo, Salesforce Agentforce or Copilot in Dynamics 365; and models called through APIs from their own tools. Most end up with all three, and the governance question is the same for each: what data goes where, who can use it, and who is accountable for the output.
Great Minds is a Sydney-based systems integrator. We help you pick use cases worth doing, set up the tools with the right data controls, connect AI to the systems and content it needs, and measure the result. If you're choosing between assistants, see Claude and ChatGPT for enterprise. If you want AI that takes actions across systems rather than producing content, see agentic AI.
Capabilities
What a generative AI engagement covers
Adoption is rarely the problem now; getting measurable value is. These four pieces of work focus on a few processes, connect AI to the data those processes run on, and measure the result in business terms.
01Knowledge assistantsAssistants that answer questions from your policies, product information and knowledge base, using retrieval over your own content and respecting each user's existing permissions.
02Content generation and automationDrafting, summarising, translating and extracting data inside the tools people already use, such as Confluence, Jira, the CRM and Microsoft 365, with a person reviewing output before it's relied on.
03AI governance and ethicsAn AI policy covering approved tools, what data may go into each, human review and incident reporting, aligned with the federal Guidance for AI Adoption and the Privacy Act.Partnerships
AI platforms we implement and connect
We work with the AI built into Atlassian, Google, Microsoft and Salesforce, and with models from providers such as OpenAI, Anthropic and Google through their APIs, including through Microsoft's and Google's cloud platforms.
RovoAtlassian's AI across Jira, Confluence and Jira Service Management: search, chat and agents working from your Atlassian content and connected apps.
Google GeminiGoogle's Gemini models and Gemini in Google Workspace, set up with data controls and connected to the systems your teams use.
Salesforce Agentforce and Einstein AISalesforce's AI for sales and service, grounded in your CRM data and governed by the Einstein Trust Layer.How we deliver
From first use case to measured results
We list the AI tools already in use across your organisation, approved or not, and the data going into each. Then we pick one or two processes where AI could save real time, ideally where it needs data from more than one system, and set a baseline to measure against.
Single sign-on, data retention settings, connectors limited to approved sources, and an AI policy people can follow. We decide with you which data each tool may see, and train staff on what to check before relying on an output.
Retrieval over your content, prompts and the integrations that put AI inside the tools people use. Each use case is tested against a set of 30 to 100 real examples with known good answers before rollout, rather than against vendor benchmarks. AI agents speed up the build; our consultants own the design and sign off every change.
Results are compared with the baseline in business terms, such as time saved or cases resolved, not just usage. What works is extended to more teams; what doesn't is stopped.
Australian considerations
What Australian organisations need to check
Australia has no standalone AI Act. The National AI Plan relies on existing, largely technology-neutral laws, so privacy, consumer and anti-discrimination law all apply to generative AI. The National AI Centre's Guidance for AI Adoption sets out six essential practices: decide who is accountable, understand impacts, measure and manage risks, share essential information, test and monitor, and maintain human control.
The OAIC's guidance on commercially available AI products says privacy obligations apply to personal information entered into an AI tool and to any output that contains it. As best practice, it recommends not entering personal information, especially sensitive information, into publicly available generative AI tools. Check each product's terms on training, retention and where prompts are processed: for some products, where data is stored and where it's processed differ. Our guide to OpenAI for enterprise explains what Australian data residency covers.
If generative AI makes, or substantially helps make, decisions that significantly affect people, the Privacy Act's automated decision rules apply from 10 December 2026. See automated decisions and the Privacy Act.
FAQ
Generative AI questions
What's the difference between generative AI and agentic AI?
Generative AI produces content: drafts, summaries, answers and translations, usually for a person to use. Agentic AI uses the same models to plan and take actions across systems, such as updating records or raising requests, with guardrails and approvals. Many projects start with generative use cases and add agents once the data and controls are in place. See our agentic AI page.
Will our data be used to train the models?
It depends on the product and plan. For business plans of the major assistants, such as ChatGPT Business and Enterprise, customer data isn't used for training by default, and Atlassian says Rovo prompts aren't used to improve the models. Free and consumer tools can differ, so check the terms of each product before staff use it.
Which generative AI tool should we choose?
Start from the platforms you already run and the tasks you want to improve. Teams on Atlassian get Rovo, Microsoft 365 organisations have Copilot, and Google Workspace has Gemini. For a general-purpose assistant, compare options against your own tasks. Our Claude and ChatGPT for enterprise page compares the two main ones.
Can generative AI use our internal documents safely?
Yes, if it retrieves content through the same permissions your staff have, so people only get answers from documents they're allowed to see. Fix over-shared folders and spaces first, because AI makes existing permission problems easier to find.
How do we stop it making things up?
Ground answers in your own content and show the sources, test each use case against real examples with known answers, use the smallest model that passes, and keep a person reviewing output wherever an error would matter. No setting removes the risk entirely, which is why human review stays in the design.
Do we need an AI policy?
Yes, and a short one is better than none. It should list approved tools, the data that may go into each, who reviews outputs before they're relied on, and how incidents are reported. We align it with the Guidance for AI Adoption and your privacy obligations.
People
AI consultants, solution architects and integration engineers
Our consultants work with business and technology teams to choose use cases, set up tools with the right controls and connect AI to your systems and content. They stay accountable for the result, using AI agents in their own delivery work under their review.
Case studies
Data and AI in practice
Generative AI depends on connected, well-governed data. These projects show the integration and data work behind it.
One asset library for APAC, EMEA and North AmericaModernising creative storage and operations for a global retailerEnabling global brand consistency through enterprise Digital Asset Management
4x faster campaign execution, 1M+ contacts migratedSalesforce Marketing Cloud to Klaviyo migration with 1M+ email marketing recipientsUnlocking personalised growth at scale through Klaviyo implementation and data migration
Segmented, personalised email for 1.4M+ recipientsSegmented, personalised email for 1.4M+ Myer customersMyer brought us in to fix segmentation and personalisation in an email programme reaching 1.4M+ recipients, with 11.2M+ emails sent.Insights
Guides to generative AI
AI in customer acquisition: what works for Australian marketersWhere AI genuinely improves customer acquisition, why connected first-party data comes first, and the Australian privacy and spam rules CMOs must plan for.
Migrating from Salesforce Marketing Cloud to Klaviyo: lessons from 1M+ contactsHow to move from Salesforce Marketing Cloud to Klaviyo: what the integration moves, Spam Act consent, active-profile costs, warming and cutover.
What is a CDP (customer data platform)? How it differs from a CRMA CDP unifies customer data from every system into one profile other tools can use.