I've been thinking about where AI is headed, and I believe we're approaching a significant shift in how organizations work. Its funny because we've said this before ... about two years ago, and a major shift did happen (ChatGPT going viral and a surge in AI interest, adoption, and funding). But I think we're on the cusp of another one. I believe this year we'll see AI agents taking on roles that look a lot like traditional employees. We've seen glimpses of this already (see footer for examples) but we'll see accelerating adoption this year.
This is driven by rapid improvements in AI model capabilities, better techniques for using these models effectively, and the development of practical infrastructure like databases, SDKs, and tooling.
I see this changing how humans work in interesting ways. Instead of doing most of the hands-on work, I think we'll shift toward being coordinators and strategists. Our value won't come from executing tasks, but from our ability to set direction, delegate to AI agents, and communicate our objectives clearly (being able to write good prompts is a skill and will become more important!). It's a big change in how we think about white-collar work. We'll take on more of a CEO role where we'll be responsible for prioritizing tasks and evaluating outputs, which will be accomplished by managing a team of AI agents to achieve our goals.
Organizations can already do more with less with the AI breakthroughs we've seen in the last year, and this will only accelerate. Entrepreneurs will find it easier to start companies, and initial products will launch with higher quality.
For AI agents to become mainstream and for them to become legitimate employees, we still need to solve some technical challenges. Agent operations need to become smoother, integration with existing systems needs work better, and the models themselves need to get faster and cheaper. Memory and context management will also need significant improvement.
These are just my predictions based on current trends. The AI world is notoriously hard to predict. What I'm certain about though is that the relationship between humans and AI in the workplace is changing – I'm just trying to understand how.
Glimpses of Agents Running Loose
- Tobi Lutke's (Shopify CEO) talk about how agents autonomously prepared his talk for him
- Claude Computer Use - Give an agent a computer and let it autonomously complete a task
- Dharmesh Shah (Hubspot CTO) talking about how agents will force us from Softare as a Service (SaaS) to Results as a Service (RaaS). Also check out agent.ai that Dharmesh built and does the very thing he's talking about
Want to chat? Shoot me an email at mathusans52@gmail.com or dm me on X.