AI Agents & The Usual Tech Hype

Right. So, the next big thing is apparently ‘agentic AI’. Honestly, it sounds like something out of a particularly ambitious sci-fi film. Apparently, it’s AI that doesn’t just generate things (like those chatbots everyone was briefly obsessed with) but actually does things. Like, completes tasks. On its own. Which, let’s be honest, is either incredibly exciting or the first step towards a robot uprising. I’m leaning towards the former, but only because I’ve invested a small, possibly foolish, amount of money in the sector.

The problem, as always, is ‘hallucinations’. Not the fun kind with sparkly unicorns, but the kind where the AI just… makes things up. With a chatbot, you can usually spot the error. You think, “Hmm, that doesn’t sound quite right,” and Google it. But if this AI is running around making decisions for a business, well, that’s slightly more problematic. It’s like letting a toddler loose in a china shop, except the china shop is your profit margin. And there are multiple toddlers, from multiple vendors, all running around at once. It’s chaos, pure chaos.

I’ve been trying to keep a log of my feelings about all this, it helps to stay grounded. It looks something like this:

  • Units of Cryptocurrency Lost: 2 (so far)
  • Hours Spent Watching Charts: 11 (and counting)
  • Number of Panicked Texts to Friends: 18 (mostly asking if they’ve sold everything)

The Orchestration Illusion?

Apparently, there’s a company called UiPath (PATH 0.17%) that thinks it can wrangle all this chaos. They used to do ‘robotic process automation’ – basically, automating simple tasks like data entry. Very sensible. Very… boring. Now they’re trying to be the ‘Switzerland’ of AI agent management, letting you govern all these different AI agents from different vendors. It sounds… ambitious. And Switzerland is expensive, so I’m anticipating high management fees.

Loading widget...

Their new platform, ‘Maestro’, is supposed to assign tasks to humans, software bots, and AI agents. Which is… logical, I suppose. Software bots are cheaper, so why use an expensive AI agent for something a simple bot can handle? It’s like using a Rolls Royce to pick up the dry cleaning. It’s technically possible, but deeply, deeply unwise.

Look, UiPath might be onto something. They’re at the beginning of this journey, and their revenue growth is starting to accelerate. But let’s be realistic. This is still the Wild West of AI. There will be failures. There will be hype. And there will be a lot of people losing a lot of money. (Possibly me.)

The stock is ‘cheap’, apparently. A forward price-to-sales multiple of below 4.5 and a forward P/E ratio of 19. Which sounds good, until you remember that ‘cheap’ stocks are often ‘cheap’ for a reason. Still, I’ve allocated a small percentage of my portfolio. I’m telling myself it’s ‘research’. My therapist is suggesting something else.

Will UiPath define the next chapter of the AI market? Possibly. Probably not. But it’s certainly a story to watch. Mostly so I can tell everyone I said so when it inevitably goes one way or another. Current emotional state: cautiously pessimistic. And slightly caffeinated.

Read More

2026-01-23 21:43