Writing / Hunter Harris
Nobody Cares About Your Agent
As I’ve mentioned before, I talk to a lot of founders. Agents are the hot hotness. They have been hot hotness for years now. It felt like personal agents had a moment back when OpenClaw released, and recently we’ve seen a “uplevel” to these agents with Instinct, Muse, Dot, and a few other solutions.
This is a competitive space. Many founders are moving into it. Here’s the problem: 90% of those founders are wrong about the opportunity agents provide.
Bring Me Problems, Not Solutions
This is the crux of the misunderstanding. An agent is a solution to a problem. Unfortunately, no one cares about solutions, they care about their problem.
I have particular goals for my business. I want to achieve them at most-any cost (some caveats apply!).
I do not care if the “how” is duct tape and chicken wire. I do not care if the solution involves me messaging my customers with carrier pigeons. I do not care if the best solution involves writing postcards to my customer. I also do not care if the best solution uses cutting edge technology. In fact, the list of things I do not care about approaches infinity!
The output, the how, the details? They just don’t really matter. Technology is not interesting or exciting in and of itself (on a business level! This is certainly not true on a personal level) - we care about being able to do more, to do the same thing more cheaply/faster/better.
Let’s be real, AI is powerful, but AI scales what exists. Weak foundations, weak scaling.
Firm Foundations
AI has data requirements.
It doesn’t matter if your agent is incredibly cool or powerful - bad data in, bad results out. MOST people and companies don’t have good data fundamentals. So if you want to sell an agent, either: you provide a LENGTHY onboarding (offloading cost onto the customer), or you sell white glove onboarding (or, this can be the meat of a service offering. Just getting people recording their calls has huge ROI now!). Someone’s going to pay for it. If it’s expensive and hard, customers will churn. If it’s expensive and easy, that’s a no-nonsense tradeoff.
Service Enhanced Software
AI is the least interesting part of an offering - it is Software, but what about the Service?
Services are way more tantalizing at this stage - it’s the software enhancing the service, not the software being provided as a service. Traditional SaaS requires huge product investment, after which the product scales, right? But right now, 6 months from now (more likely 3), the software you’ve built will be dated. Will you pay cost to maintain it to scale it? Or is it better to just throw it out and refocus elsewhere?
Services are flexible. If the meat of the offering is in the service, and AI is just accelerating it, the software is fungible. Instead of trapping yourself in months-long build cycles, you can just… throw stuff away. New opportunity? New offering. Some of the software will transfer, other parts won’t make the cut, and that is fine. Service focused offerings actually have a chance to survive the pace of innovation here, whereas lengthy build runs will struggle more.
Seriously, the technology is the least interesting part of most of these offers. And that’s why the moat is interesting.
Yes, the software can be copied or replaced. But the software is just accelerating an offer’s ROI, or making it more scalable. If you copy it, you still aren’t grabbing the expertise. Vibe coded replacements just don’t matter for these offerings.
The opportunity is real here. But you have to orient yourself towards the current opportunity, not the historical one. This landscape will change! I expect we’ll eventually see some durable products shake out of this market. But we’re not there, and we have to survive till we reach that state. Which means staying nimble and adapting.