In 2027 we hand AI whole chunks of work
In 2027, we'll be delegating whole chunks of work to AI.
Not just a task, not just talking to it about a topic, but really handing over a whole job and letting it go ahead and do it.
I'm seeing this already happening.
Think about an AI assistant that organises your inbox, tags emails by topic or urgency and files them away. Not just a summary or one draft, an actual helper in your inbox moving things around.
Or AI that prepares activity statements or GST for a group of clients, bringing the reports together, checking last quarter for instalment adjustments and getting everything ready in one place for your review.
Or onboarding a new client. AI organises the documents, prepares letters, attaches them to draft emails and starts enters information into different systems. The whole process of setting up a client or project.
The difference is that we're handing over the work as a whole, not task by task. We're not sitting there saying, "Now get me the report, and now put it into a tracker, and now..." We're keeping the whole task together.
AI has become very, very good at that. It plans, it reasons, it loops and checks itself and it hands out high quality, ready work.
Let me show you how I know
I know this is what's coming in 2027 because I look at what the innovators and the early adopters are doing, and I know that's what we're all going to do very soon.
To give that real substance, I got my AI to go through all the posts in our AI for Accountants club's forum, the transcripts of our meetups, the meetup chats, and all the transcripts of my one-on-ones with club members, and I got it to map all the use cases.
I organised them into six levels (you can see a visual in the video up the top).
Set up. This is for people who are new. We just make sure they've got the right policy and the right licences in place, and they understand what happens with their data.
Ask. This is where we chat with AI. We upload, we download. This is very 2025.
Work. We bring AI into our workflow, using AI inside Word, Excel, PowerPoint or Outlook. It doesn't matter if we use Claude, ChatGPT, Copilot or Gemini, they all do the same thing.
Reuse. This is where we see a massive cluster of use cases. We're building agents or creating skills. Essentially saved prompts with some knowledge that act as playbooks and come in whenever we need to do this task.
This is where a lot of the work is happening right now. It's today's mainstream, or early majority if we think of the technology adoption curve.
Delegate. We're giving AI an outcome to work towards and letting it work through the steps. It already knows the playbook, and now it can gather what it needs, do the work and bring it back for us to review. We're still setting the boundaries and making the decisions, but we're not prompting every step.
This is the inbox helper, the GST preparation and the client onboarding I mentioned earlier all sit in this delegate level.
Build. The innovators are building dashboards and small apps, and sometimes even replacing some of the apps they're paying for with something AI builds for them. This is what I did with the use cases dashboard here.
Where the majority is today
44% of the use cases shared in our club sit in Reuse. This is what’s mainstream today.
The next biggest level is Delegate. This is what mainstream will be soon.
Your next step
In your favourite AI, go and look for those areas where you can create skills or agents. This is how you teach it the playbooks of how you work and make it all reusable and consistent across your team.
And then go and identify that Cowork area. That's where you have all the skills and playbooks you taught it, and all your connectors to other systems. And you can set up tasks to happen on a schedule or on a trigger, and that's where we create that magic of delegating to the AI.
We’ve got this.
--
Inbal Rodnay
Guiding Firms in AI Adoption and Automation
Keynote speaker | AI Workshops | Executive briefings | The Tech Savvy Firm
When you are ready, here is how Inbal can help:

