Artificial intelligence has captured the attention of almost every industry.
Business owners are experimenting with ChatGPT, Microsoft Copilot, Claude and a growing list of other platforms. Teams are using them to rewrite emails, summarise documents, brainstorm ideas and find information more quickly.
Those uses can be helpful, but they often leave businesses asking an important question:
What is AI actually doing to improve the way we operate?
That was one of the central themes of Michael Brook’s presentation at our recent Geelong Cyber Sip event.
Michael focused on the next stage of business AI: agents that do more than respond to an occasional prompt. These tools can work across the systems a business already uses, find relevant information and complete parts of a process with far less manual effort.
It was an informative and practical session that showed how quickly AI is moving beyond novelty.
This is also one of the reasons we run Cyber Sip events.
Much like our Lunch & Learn sessions, Cyber Sip is about bringing useful and current information directly to the business communities we support. Technology is moving too quickly for organisations to rely on what they heard several years ago.
We want business owners to understand what is becoming possible, what the risks are and how they can make informed decisions about what comes next.
Michael’s presentation did exactly that.
Most businesses do not have an effort problem
Teams are already working hard.
The difficulty is that a significant amount of their day can disappear into tasks that support the work without creating much value themselves.
People search for old emails.
They look through folders for the latest version of a document.
They take notes during meetings and then rewrite those notes into an update.
They manually collect contact information from messages, calendar invitations and spreadsheets.
They answer the same internal questions repeatedly.
None of these tasks is necessarily difficult.
They are simply time-consuming.
When they are repeated across a team, every week of the year, they take an enormous amount of attention away from customers, planning and the work people were actually employed to do.
This is where AI agents begin to change the conversation.
Instead of waiting for someone to open a tool and ask a question, an agent can be designed to carry out a useful task as part of the normal flow of work.
What makes an AI agent different?
Most people’s first experience with generative AI involves a chat window.
They enter a question and receive an answer.
That can be useful, but the interaction remains separate from the rest of the business. The person still needs to locate the right information, copy it into the tool and decide what to do with the response.
An AI agent can go further.
It can be given access to approved information across systems such as Outlook, Teams, SharePoint and OneDrive. It can retrieve information from those systems, interpret the context and then assist with a particular action.
That might involve identifying missed follow-ups, preparing a meeting summary or building a first draft of a business document.
The agent is not simply answering a general question from the internet.
It is working with the information and processes relevant to that organisation.
Michael described this as the point where AI begins to move from something interesting into something genuinely useful.
Starting every day with the work that matters
One of Michael’s examples was a simple agent he uses to review emails from the previous day.
Rather than manually searching through his inbox and trying to remember which messages still require action, the agent identifies the emails he has not followed up and presents a list of what needs attention.
For someone working in sales, account management or business development, that can be extremely valuable.
Important messages are easy to lose during a busy day.
A customer asks for additional information. A prospective client mentions they will be ready to talk next week. A colleague requests input on a proposal.
The email is read, but something more urgent happens before a response is sent.
Several days later, the person suddenly remembers it.
An agent can provide another layer of protection against that kind of oversight.
It is not replacing the relationship or writing every response without supervision. It is making sure the opportunity does not disappear into a crowded inbox.
That is a practical use of AI.
It solves a real problem and gives the person a clearer starting point for the day.
Turning meetings into action
Meetings are another area where useful information is frequently lost.
Many businesses now record or transcribe online meetings. The transcript contains the discussion, decisions and action items, but someone still has to review it and turn it into something useful.
An agent can assist with that process.
After the meeting, it can identify the major decisions, collect the tasks assigned to each person and prepare a summary for the attendees.
It could also create tasks from the action items or draft a follow-up email for someone to review before sending.
This reduces the administrative work after the meeting and makes it less likely that an agreed action will be forgotten.
The benefit is not simply that the summary is created more quickly.
It also creates greater consistency.
Everyone receives the same record of what was discussed, what was decided and what happens next.
Finding information without knowing exactly where it is
Most businesses have more information than they know how to use.
Documents sit in SharePoint.
Conversations are spread across Teams.
Emails remain in individual inboxes.
Files are stored in OneDrive, shared folders and other systems.
The information exists, but locating it can take time, particularly when the person does not remember the exact name, date or location.
Michael described how AI can use context to help find that information.
Instead of searching for a specific subject line or filename, someone may be able to describe what the conversation was about.
They might remember that they spoke with a customer about a particular issue sometime last year, but not remember when the email was sent.
An AI tool connected to the organisation’s information can use that context to find the relevant communication.
That is a significant change from traditional search.
The user no longer needs to remember the exact words used in the document. They can search based on meaning.
For employees who spend a large part of their day navigating information, the time saving can be substantial.
Building useful business tools without becoming a developer
One of Michael’s most compelling demonstrations involved creating a data security workshop.
A colleague needed a structured engagement that could help assess a customer’s security maturity and support a useful business conversation.
Traditionally, building that resource could involve hours of research and preparation.
Questions would need to be developed.
A scoring framework would have to be created.
The results would need to be connected to recommendations and an action plan.
Michael gave Microsoft’s Cowork capability a brief description of what he wanted.
From that short instruction, the system created a set of questions, the scoring logic, a summary process, a maturity structure and an action plan. It then produced an Excel scorecard that could be used in the workshop.
The result still required review.
Michael was clear that generative AI should not be expected to be completely accurate. Someone with subject-matter knowledge needs to assess what has been produced and correct anything that does not make sense.
But the AI completed a large proportion of the initial work in minutes.
That is where the real opportunity sits.
It does not remove the need for expertise.
It allows the expert to begin much further along.
Collecting business data that is already there
Michael also demonstrated an agent used to build a contact list from the systems he already worked in.
He asked it to review Teams, calendar invitations and his inbox, identify external business contacts and place their details into an Excel spreadsheet.
The system collected names, email addresses and company information. It recognised common internal or partner organisations and asked whether they should be excluded.
Once the information was in Excel, Copilot helped identify and remove duplicate contacts without Michael needing to manually specify the exact spreadsheet formula or column.
The task resulted in a structured list containing hundreds of contacts.
The value of that example is not the contact list itself.
It is the fact that the information already existed across the business.
The time-consuming part was bringing it together.
Many organisations have similar opportunities hiding in plain sight.
They already have the data required to improve customer service, sales follow-up, reporting or internal knowledge.
The challenge is retrieving and organising it.
The best agents solve small, recurring problems
The temptation with AI is to begin with an enormous transformation project.
Businesses imagine completely autonomous systems that run departments and make major decisions.
That is not always the best starting point.
The most useful initial opportunities are often smaller.
What does someone on the team do repeatedly?
Where are people searching for the same information?
Which follow-ups are easy to miss?
What needs to be summarised, sorted or transferred from one system to another?
An agent that saves one person 20 minutes a day may not sound revolutionary.
Across a full year, that is a meaningful amount of time.
If the same agent helps an entire team, the value increases quickly.
The best early use cases tend to be clear, repetitive and easy to review.
They create a visible benefit without requiring the business to redesign everything at once.
AI still needs people who understand the work
Michael did not suggest that AI agents can operate successfully without human involvement.
The person using the tool still needs to understand the subject.
They need to recognise when a result is wrong, incomplete or inappropriate.
An agent can create a strong first draft, but someone needs to review it.
It can collect information, but a person needs to decide what should be done with it.
It can suggest an action, but the business needs to determine which decisions require human approval.
This is why AI capability is becoming an increasingly valuable workplace skill.
Knowing how to use a particular software package has traditionally been part of many job descriptions.
Businesses are now beginning to look for people who understand how to work effectively with AI.
Michael shared the example of an organisation that gives new employees time during their first week to build an agent.
The purpose is not to turn every new starter into a programmer.
It is to encourage them to identify a problem, understand the available information and think about how AI might improve the process.
That mindset will become increasingly valuable.
Free time does not remain free for long
Michael made an interesting observation about major technological changes throughout history.
New inventions frequently save people time.
That time is quickly filled.
The smartphone gave people access to information and communication from almost anywhere, but it did not make people feel less busy.
AI is likely to follow a similar pattern.
The purpose is not necessarily to create an organisation where everyone has nothing to do.
It is to reduce the time spent on work that does not require a person’s experience, judgement or relationships.
That allows people to focus on more meaningful activities.
A customer service employee can spend less time searching for information and more time helping the customer.
A salesperson can spend less time assembling contact lists and more time building relationships.
A manager can spend less time rewriting meeting notes and more time making decisions.
The business remains busy.
The goal is for that time to be used more effectively.
Why this matters to us
At eManaged, we believe technology should create a clear benefit for the business using it.
AI should not be adopted simply because it is new or because everyone else is talking about it.
It should solve a genuine problem.
It should remove unnecessary friction, help people find information and make useful work easier to complete.
Michael Brook’s presentation at our Geelong Cyber Sip event showed what that can look like in practice.
We were incredibly pleased to have him share these examples with the local business community. Rather than speaking only about what AI may be able to do in the future, Michael demonstrated tasks that can already be improved today.
That is why events like Cyber Sip are so important to us.
They help business owners stay informed about a rapidly changing technology landscape and give them practical ideas they can take back into their organisations.
The next step is not to build dozens of agents overnight.
It is to identify one recurring task that consumes more time than it should and explore whether AI could make it easier.
Talk to eManaged about where AI and automation could create practical productivity gains in your business.