Issue #7 · March 26, 2025 · 4 min read
Welcome to the new & improved Building Better.
I’m breaking things into sections to give you busy people some time back in your day.
This week we’re watching how AI agents are quietly transforming industries while we're still debating if Facebook is worth saving, or it’s gone to the bots. The gap between public perception and industry reality has rarely been this wide.
ANALYSIS
The Quiet Revolution of AI Agents
I set up what I thought was a simple automation task.
Step 1 - once a week have an AI scan my emails for receipts.
Step 2 - upload them to Google Drive.
Step 3 - reconcile them in my accounting software.
What I didn't expect was to receive an email from "myself" saying it wasn’t able to find any receipts.
I never programmed it to email me updates. It just... did.
When most people think of AI today, they imagine tools generating blog posts or images. But content generation is just a tiny drop of what's possible.
The real revolution is in AI agents. Autonomous systems that can reason, make decisions, and complete entire tasks independently.
This goes beyond just building better tools, we’re sort of making digital collaborators.
For example Westpac, one of Australia's largest banks. They've implemented an AI agent to help programmers shift computing code from legacy systems to cloud-based ones. Work that previously took a human engineer six days can now be completed by an AI agent in one hour.
The difference is not doing the same things faster, it's doing things that were previously impossible due to human limitations.
As these systems become more sophisticated, we'll increasingly interact with them as we would with colleagues. Through email, chat platforms, and collaborative workspaces, maybe even phone calls.
Are we ready for a world where our technology starts taking initiative? Whether we're ready or not, that world is already here.
SIGNALS
1. The Agent Marketplace Is Coming
Agent.ai has emerged as the first marketplace where people will be able to buy and sell AI agents created by other users.
Started by the Dharmesh Shah, co-founder of Hubspot. Agent.AI is like a merge between an app store and Fiverr. You can use skills for hire, but packaged like apps.
My prediction is that small businesses that could never afford to build custom AI solutions will simply be able to purchase pre-built agents, the same way they might download apps today.
2. Agents In Regulated Industries
Melbourne based startup Minikai just raised its seed round to revolutionise how care providers operate in highly regulated sectors through AI-powered automation.
Their platform gives each patient a dedicated AI agent that understands their specific needs and history, while transforming the workflow of frontline care providers by automating administrative work and compliance reporting.
Administrative burden is crushing (compliance teams spend over 50% of their time on audits and paperwork) and stakes are high (fines exceeding $1.5 million for compliance breaches). This burdensome work can now be handled more efficiently and accurately, which allows care staff to prioritise their patients wellbeing.
3. Entry-Level Roles Will Change First
Virtual Assistants and data entry positions are likely to see the most immediate impact. According to Taskdrive there are over 40 million Virtual Assistants globally as of 2024.
The critical question: what happens to people in these more administration based roles? What new skill sets do they need to develop? How does career progression change when traditional starting points disappear?
FINDS
🔍 Collaborative Intelligence - HBR's piece on human-AI collaboration feels more relevant today than when it was published in 2018.
💼 Minikai's Seed Round - Worth reading Tidal Ventures' investment notes on why they led this round.
🎧 Approximately Correct - This episode discusses how human-AI collaboration can enhance decision-making and performance.
📱 n8n - A no-code tool for creating simple AI agents. If you're looking to experiment, this is a good place to start.
If you’re already automating tedious tasks, I'd love to hear your thoughts.
Until next week,
Alex
P.S. If you found this valuable, consider sharing it with someone who's trying to make sense of these shifts. And if you're new here, subscribe to get these insights in your inbox every week.
In project management land, the AI smarts in tools like Asana and Notion are increasing discoverability of features (eg. Notion AI suggested adding a 'shoot location' custom field when I asked it to setup a basic photography project tracker), but you have to be wary of accepting all suggestions and bloating your data 😬
Great post! Thank you for sharing and inspiring us to explore them!