I spend a lot of my week going through software. Some of it is already live, but a lot of it I see through my work with AppSumo while the product is still being evaluated, tested, fixed, repositioned or getting ready to launch. It is a slightly weird point in a product’s life because most of the obvious work has been done. The thing usually works. The more interesting question is whether it works far enough.

I have noticed myself asking the same annoying question more and more while testing products: okay… and then what? You found me a lead. Great, and then what? You told me traffic dropped. You identified an opportunity. You generated a beautiful report explaining exactly what happened. All useful things, but what am I supposed to do next?

It sounds almost stupid when I write it like that, but I think there is an important product question hiding inside it. A surprising number of products get the customer very close to the thing they actually wanted and then stop. Sometimes that is exactly where the next feature, integration or even the real product opportunity is hiding.

Postbeam is probably the cleanest example because I wrote down the gap almost immediately when I first looked at it. What interested me was that it wasn’t really another “AI writes LinkedIn posts” product. There are enough of those. Postbeam was trying to connect team posting, approvals, analytics and actual engagement signals into something much closer to a warm prospecting workflow.

Someone likes your post, comments, or keeps showing up around your content. Postbeam can identify that person, compare them against your ICP and surface the people who may actually be worth talking to. That is considerably more interesting to me than generating another LinkedIn post, because now the software is starting to connect content to an actual business outcome.

But when I wrote my Field Note about Postbeam, the part that immediately jumped out at me was what happened after the warm lead appeared. I wanted deeper CRM continuity, stronger follow-up, better attribution and more of a closed loop between “this person is interested” and whatever the sales team needs to do next. That was the part that could move Postbeam from being a clever LinkedIn tool into something closer to a real GTM system.

I went back to Postbeam while putting this issue together and some of that gap has already started closing. Warm leads can now move into HubSpot, Salesforce or CSV, it can prepare an opener based on the post someone actually engaged with, and there are follow-up campaigns around those signals too. I don’t say that as some great “look, I predicted the future” moment. The direction simply makes sense. Detecting the lead is useful. Helping move that lead toward an actual conversation is much closer to the outcome the customer wanted in the first place.

Then there is Sleek Analytics. I first wrote about Sleek in my Observation Deck because I liked the idea behind it immediately. Analytics software has developed a strange talent for giving people enormous amounts of information while somehow making them less certain about what is going on. I have used Google Analytics for years and there are still moments when I open it and feel like I accidentally walked into somebody else’s cockpit.

Sleek goes the other way. Show me where people came from, what they did, what pages matter and what is happening on the site without forcing me to become an analytics consultant first. That simplicity is useful, but the same question eventually appears. Knowing that 400 people visited a page is information. Knowing that one traffic source converts much better than another is more useful. Knowing which page, campaign or referrer actually produced revenue gets considerably closer to what the founder was probably trying to understand in the first place.

Sleek has been moving in that direction with revenue attribution and the ability to ask questions about the data in normal language. That’s the part that interests me. I don’t really need another dashboard. I need software that reduces the distance between something happening and me knowing what decision to make because it happened.

Airtop comes at the same problem from a completely different direction. I first included it in my Field Notes because the basic idea is simple and slightly nuts: give the AI a browser.

That means an agent can log into a site, click through a workflow, fill forms, collect information, download reports, update records and do many of the boring browser things that normally require a human sitting there moving a mouse around. APIs are wonderful when they exist and expose everything you need. Quite often they don’t. The browser is still the ugly universal API sitting in front of us.

What interests me about Airtop isn’t browser automation by itself. It is what it represents. For years we have built software that tells us things. AI made that dramatically better. Ask a question, get an analysis. Feed it data, get a recommendation. Give it a problem, get suggested next steps. Now the software can increasingly take some of those next steps too.

That feels like an important progression. The answer box doesn’t just get smarter. The distance between the answer and the work getting done starts shrinking.

Postbeam, Sleek and Airtop are three very different products. One is around LinkedIn prospecting, one around website analytics and one around browser automation. I wouldn’t normally put them beside each other, but when you see enough software you start noticing the same product decisions appearing in completely unrelated categories.

I think these three are exposing the same SaaS workflow problem: where does the customer’s definition of “done” actually sit? Because it may be further along than where your current feature stops.

I see versions of this all the time while evaluating SaaS. A product surfaces an SEO problem but doesn’t help fix it. Another identifies leads but leaves all the follow-up somewhere else. Another creates a strategy document but doesn’t connect the strategy to execution. Another produces an AI analysis that is genuinely impressive and then gives the user a Copy button.

Copy it where? To do what? Which other software do I now need to open? How many steps are still left before I get the outcome I thought I was buying?

This does not mean every product should become an all-in-one monster. Please don’t. I have tested enough giant software platforms that somehow want to be your CRM, email tool, project manager, social network, accountant and therapist at the same time. Focus is still a feature.

But there is a difference between keeping a product focused and stopping the SaaS workflow too early. The trick is figuring out which side you’re on. Sometimes the handoff is perfectly reasonable. Other times you are handing the customer off one step before the thing they actually care about happens.

Putting these notes together reminded me that I have already been writing versions of this idea elsewhere on the site. In my Growth Hacking Tools guide, I make the case that growth software should be connected to a specific problem, a hypothesis, something measurable and a workflow you can repeat. Otherwise it becomes incredibly easy to create a lot of activity without learning very much.

You can get more visitors, send more outreach, run more experiments and fill more dashboards. None of those things automatically mean you made a better decision. Did the user activate? Did the lead turn into a conversation? Did revenue move? Did you learn something useful enough to repeat? Those are much closer to the questions I care about.

Read the Growth Hacking Tools guide here.

The same principle appears in my Marketing Funnel Design and Building Tools guide. A funnel isn’t a landing page. It is the complete path from the traffic source through the page, form, email, checkout and follow-up. A gorgeous landing page with broken follow-up is still a broken funnel, which is why I keep coming back to the full workflow instead of whichever individual step happens to have the nicest software around it.

I noticed the same idea again from another direction in my GEO tools guide. A growing number of products can tell you whether ChatGPT, Claude, Perplexity and the other answer engines mention your company. That’s useful, but the products that interest me more are the ones that start answering the next questions too. Which page needs work? Which questions are buyers asking that you haven’t answered? Why does a competitor keep appearing where you don’t? What should you actually change?

Different categories, different software, same question: and then what?

If I were building SaaS today, I’d ask one question about every important feature: and then what? If my product finds the lead, spots the problem or generates the answer but leaves the customer to complete the valuable next step somewhere else, I’d look very hard at that handoff. That’s the part of the SaaS workflow I’d be looking at hardest.

I think this matters even more as AI makes the intelligence layer cheaper. Another product can probably summarize the same meeting. Another can analyze the same spreadsheet. Someone else can probably generate similar text using Claude, GPT or whatever model appears next Tuesday. The intelligence itself is becoming easier to reproduce.

What becomes more interesting is everything around it. What context does the product already have before the AI is called? What happens after it replies? Which systems does the result connect to? How many decisions, copy-pastes and annoying little steps disappear before the customer gets to done?

That is where I think a lot of the product opportunity is moving. Not necessarily toward adding more AI, but toward finishing more of the job.

And sometimes the best place to look for the next feature isn’t inside your product at all.

Look at the first thing your customer does after leaving it.

There might be something there.

More soon,

Abe

Whenever you’re ready, there are 2 ways I can help you:

1. Get your app or website reviewed by me.
2. Work with me 1:1. After evaluating more than 1,000 SaaS products, I know what makes buyers pick one over another. I help you get those differentiators right and make a better mousetrap, the kind that sells. Apply!

Join 10K+ readers of Before You Build, the newsletter for Strategic SaaS Founders who want to decide what is worth building, make it meaningfully different, and turn it into something people want to buy.

Submit

No spam. Just practical lessons from evaluating hundreds of SaaS products and helping founders make better product decisions.