We recently rebuilt Lessie AI around a different assumption: the primary user is no longer a person clicking through a dashboard, but an agent executing tasks.
Lessie started as a pretty standard SaaS for people discovery and outbound. Search, filter, inspect, export, send emails. It worked, but the interaction model started to feel dated.
The shift isn t just AI inside SaaS. Software is becoming an execution layer for agents.
So we stopped asking how to add AI, and started asking what the product looks like if it s designed for agents first.
I run a small online laptop store in the UK and pretty much do everything myself, from SEO and product pages to fixing random site issues. Over the last year my tool stack has got out of hand, and I'm sure I'm paying for a few things I open once a month.
Trying to cut it down to stuff that genuinely saves me time. What are you paying for that you'd be annoyed to lose? And is there anything you cancelled and never missed?
When it comes to sharing complex ideas, I often rely on direct conversations with friends or partners. Misunderstandings can easily lead others in the wrong direction, wasting valuable team time. To avoid this, I focus on tools that make my thoughts clearer and more structured. For tools, I usually turn to Keynote when I need a flexible whiteboard space to lay out ideas visually. Adding elements like relationships and logical connections helps make my thoughts easier to understand. If time is short, I switch to mind maps. My go-to tool is XMind
I keep wondering whether every company is eventually going to need a company-wide AI memory.
Right now, most businesses have knowledge scattered across docs, Slack, calls, CRM notes, shared drives, and (probably most importantly) the heads of a few people who have been there forever.
That was already a problem for onboarding and operations.
I'd present roadmaps and projections like they were facts. Pricing decisions like they were calculated. Hiring timelines like they were guaranteed.
They weren't. They were structured guesses with good formatting.
The shift happened when I started saying "this is our best bet right now, here's why" instead of presenting everything as certain. Decisions got faster. The team trusted the reasoning more because it was honest.
I've been active on PH for a few weeks now, mostly commenting on launches and joining discussions.
One thing I didn't expect: the feedback here is more honest than almost anywhere else.
Not because people are harsh. But because makers here have actually shipped things. They know what a real problem looks like vs a solution looking for one.
I've found that reading comments on other people's launches teaches you more about product positioning than most courses do. You see exactly what confuses people, what excites them, and what makes them scroll past.
Agent drift is so annoying. you approve a clear outline and halfway through it rewrites a file it shouldn't even be looking at or claims a task is finished when it missed half the steps.
I started locking down access and only letting it handle small task pieces one by one instead of broad plans.
Does keeping tasks super small stop the code jumping, or do you still have to babysit every edit?