Connect with the Product Hunt community

I ve been using prototypes a little differently since I started building more with AI.

For internal reviews, I still find simple prototypes useful. They help the team understand the overall flow and discuss whether we re looking at the same problem.

But when I want to explore how an idea actually works, I often go straight to building a working version with AI.

The interesting part is that I m not necessarily trying to skip the design process.

I've been thinking about adding a weekly digest email. A short summary of what changed, what's new, what's worth looking at.

The case for it is habit. A weekly email is a rhythm, and rhythm is retention. People forget products exist, and a gentle reminder is often the difference between someone staying and someone quietly drifting away. Some of my favorite products earn their place in my week through an email I actually look forward to.

I run a tech site where we publish software reviews and comparisons, so this is something I want to get right.

A feature checklist can say two tools both support exports. But if one puts exports behind a $99 plan, that single checkmark hides most of the buying decision. The same goes for AI credits: a monthly price tells you very little without knowing what one finished task costs.

My vote is for showing the cheapest plan that can actually complete a specific job, with its limits beside the price.

Think of the last tool you tried or paid for. What did you only discover after signing up that you wish the comparison had told you?

Every one of them looked good on paper. Scalping on order flow, mean reversion, the Turtle rules across 19 markets, a Sunday gap trade in forex, funding and ML filters on top of a trend strategy.

Real data killed all of them. The one that survived is a plain trend follower on ETH: +79.5% over 6.5 years in backtest. Buy and hold made far more, but its worst drop was 79%. The bot's was 6.9%.

Thinking about shipping my next mobile app in several languages from day one instead of English first.

For those who tried it: which languages actually brought installs, and which ones weren't worth the effort? Curious whether German and French users really convert better, as people say.

I attended a VoiceAI event held by NVIDIA at their Bengaluru, India office. It was very interesting to see all the startups that presented. Adi Margolin, Product Manager at NVIDIA, added a lot of her own insights at the work being done by her team in this field. For India especially, a country known for 121 major languages and about 20,000 dialects, there's a big challenge to be overcome. However, some of the results showcased were promising, with some teams targeting low latency, while others focused on improving accuracy. Interestingly, a lot of these startups are developing voice agents for the medical industry, like for eg. answering queries by patients for prescriptions.
I would love to know what all your thoughts are. I've found this an interesting concept ever since I've had an Alexa device in my home. Do you think think that Voice AI has the potential to be the next step after chatbots?

I m making the video & static creatives myself and figuring out Ads Manager as I go. It s working, sort of, but I have no reference for what good looks like yet.

If you ve done this before: what did you get wrong in the first month that you d tell someone to just not do?

We learned the hard way that free users and paying customers want totally different things. Free users usually ask for dozens of features to avoid buying other tools while paid customers just want speed and simplicity.

After wasting months on a feature that non-paying users demanded we tagged support feedback by payment history. Focusing mostly on paid requests cut our user drop-off in half and saved us tons of dev time.

We recently learned that getting a clean HTTP 200 response doesn't mean your database actually saved the changes.

In our setup, missing security rules and webhook edge cases caused multiple user updates to fail quietly without leaving a single trace in the logs. We had to change our approach and start reading back saved records or checking affected row counts to verify every write.

We spent almost three months building and maintaining a feature that 40% of our free users asked for. But when we looked at the data, that exact feature was causing 60% of our support tickets and confusing high-paying customers.

Last month, we ripped it out completely. We lost a few vocal free users, but our main product speed improved, support volume plummeted, and paid upgrades went up.

Have you ever sunsetted a popular feature to save your main core value proposition?

We had two full-time SDRs sending 150+ cold emails a day. The pipeline looked busy, but closed deals were painfully slow, and customers acquired through cold outreach had double the churn rate.

We stopped all cold outreach and shifted that energy purely into writing teardowns and technical case studies. Our lead volume dropped initially, but our conversion rate from lead to paid tripled prospects already trusted us before the call.

Have you ever killed an active sales channel because the quality of customers was hurting your business?

i asked here yesterday how people keep their agents from forgetting between sessions. almost everyone said a file in the repo.
what i took from it: never trim the file, keep the rejected idea next to the decision, small checkpoints, and make the agent write the failing test first.
mine still re-reads everything every time, so checkpoints is what i m trying next.
what does your agent still forget even with all this?

I've been building enterprise software, and one of the harder lessons I'm learning is that a positive conversation doesn't necessarily mean you're closer to a customer.

I've spoken with leaders who recognize the problem we're addressing, can describe how it affects their teams, and are interested in seeing a solution.

But interest and adoption are two different things.

In enterprise environments, the person experiencing the problem may not control the budget. Even when there's a potential business case, there can be questions about implementation, security, organizational readiness, and competing priorities.

I still answer every support email personally. Every bug report, every billing question, every "how do I do this" and every "this is broken and I'm furious." It takes hours I don't have, and some days it's the only thing I actually get done.

I do it because the signal is irreplaceable. Support emails are the most honest feedback I'll ever get people tell me exactly where the product confuses them, exactly what broke, and exactly what they wish existed. If I hand that to someone else, I lose the pulse of my own product and start making decisions from dashboards instead of from real frustration.

A customer asked me last week for a full export of everything they've put into the product every record, every file, every note. And I froze.

My first instinct was trust. Of course they should be able to take their data. It's their data, not mine. Refusing would be exactly the kind of lock-in behavior I've always hated in other products. A product that holds your data hostage isn't a product it's a hostage situation with a login page.

We used to offer a standard 20% discount code whenever a demo prospect hesitated or asked for lower pricing. It felt like we were closing deals, but buyers kept stalling, waiting to see if they could haggle for more.

Two months ago, we instituted a strict Zero Discount Policy" across the board fixed prices for everyone. Surprisingly, our deal cycle shortened by two weeks. Buyers stopped playing negotiation games and made decisions much faster.

Have you ever stopped offering discounts and noticed that prospects actually respected your value more?

Every CRO guide told us to keep our hero section clean, concise, and minimal. We had a super short landing page with just a bold headline and a CTA button, but conversion stayed stuck at 2%.

We decided to do the exact opposite: we rebuilt the page into a 3,000-word long-form teardown showing the exact step by step workflow real screenshots and edge case limitations. Conversions jumped to 5.4%.

It turns out high-intent buyers wanted deep clarity, not slick marketing buzzwords.

We were getting plenty of signups with our 14day free trial, but 80% of users were ghosting us after day 3. They simply didn t have enough skin in the game to properly set up the product.

We decided to replace the free trial with a 30day no questions asked" money back guarantee. Signups dropped by 25%, but activation soared, and 30-day retention doubled because users were actually committed to testing it out.

Have you ever gated your product more strictly and seen better long term retention as a result?

Back when I was doing dropshipping, honestly the worst part was testing. You find a product, make the store, run ads... and a week later you find out nobody wants it. Money gone.

So how do you do it now?

• do you test before launch or just launch and pray lol

• what's the cheapest test that actually worked for you?

AI tools can now analyze years of employee performance reviews, retention metrics, and project outputs to generate hiring profiles and team structures: "hire for X trait," "keep teams under Y size."

I ve spent over a decade leading cross-functional teams, and I m skeptical. Not because the AI can t spot correlation, but because I ve seen how easily corporate data misrepresents human dynamics. A team can look high-performing on paper while burning out its best people behind closed doors. If the AI learns from that HR record, it standardizes burnout with total confidence.

Here s where I ve landed so far. I d need three things before acting on an AI-generated organizational lesson:

1. The qualitative feedback, visible. Skip-level notes, team sentiment, and anonymous feedback loops. Not just productivity charts.

2. The market context at the time. A hyper-aggressive growth playbook from a bull market will backfire in a retention-focused downturn. The lesson should state when it applies.

3. A leadership sign-off. A manager who navigated the human side confirms the nuance before it becomes an algorithmic hiring filter.

Our brain run on autopilot more often than we realise. One such behaviour is hearding, where we have the tendency to follow group or try to in a group. Imagine a few people suddenly started running. You don't know what's happening, but you might start running too, assuming they know something you don't?
Now, lets apply this into Cusomer behaviour, a D2C product goes viral and suddenly everyone wants to try it. A resturants has a queue outside it, we assume food must be good. A SaaS product gets #1 on product hunt and everyone wants to know what it does.
This i find really intriguing, becuase popularity can become a shortcut for credibility. (PS not shade to people who gets #1 on this platform a very big deal and hardwork) I'm just naturally curious on cusomer behaviour, read this paper recently in harvard business review about how fast consumer behaviour is changing in this decade.
Curious to hear what other makers think.

I tested 10 prompts like "best tool for X" in ChatGPT, Perplexity and Gemini. Our product came up in 2 of 30 answers, and a competitor with less traffic showed up in 14.

Curious how others approach this:

• Do you track AI mentions at all, or is it too early to matter?

• Has anything you changed (docs, reviews, comparison pages) actually moved the needle?

• Is "AI search visibility" a real channel yet, or mostly hype?

Been here 2 months now and something finally clicked. Half the upvotes and half the comments on any launch are people returning a favour.
You upvote mine, i comment on yours, everyone goes home happy.

Nobody says it out loud because we all benefit from it. It s like a group chat where everyone agreed not to talk about the group chat.

Doesn t make PH useless tbh. It just means rank and comment count show who has the bigger circle, not whose product is actually better.

Maybe i m wrong and i m just new here.
Would love to hear from people who ve been around longer, does this match what you ve seen or have i been reading it all wrong.

Genuine question for freelancers and small agencies.

For big projects, a contract is a no-brainer. But for a $300 logo tweak or a two-day dev fix, it can feel like overkill. Sending one can even make the client hesitate.

I ve gone both ways:

• No contract: fast start, but twice the scope crept and once the payment took 60 days.

• Contract every time: slower start, but no awkward that wasn t included conversations.