Product-hunted and not ready to launch

How we built our first bot and what we've learned from it

When you set out to build something, the end result never ends up like you imagined.

At some point, the rug will be pulled and you and your team will find yourselves scrambling, desperate to make it all work.

It’s an exhilarating, exhausting ride. And, it’s absolutely worth it.

Luis Novo, the CEO of Skore, was doing customer research in California. Everywhere he looked, Slack cropped up. Teams of all shapes and sizes were using it to communicate, share links and stay on top of key projects.

At the same time, the rest of the Skore team was struggling with a different challenge. The philosophy and idea behind Skore is to optimize knowledge flow and find better ways for teams to share, and track, knowledge.

Every day, thousands of relevant links got shared in various Slack channels but they got lost in the noise. Even finding the links was hard and time consuming!

We needed a solution.

Meeting a burning need

That weekend, Luis downloaded Botkit and built the first version of Paperbot. The idea behind the bot was simple.

It would capture the links teams shared in Slack and transform them into an email digest.

Late on Sunday night, Luis sent the rough version, along with research notes, to the rest of the team.

The excitement in the office that Monday morning was palpable. Paperbot could be the missing link.

Creation often stems from a need and Paperbot was going to meet ours. Instead of wasting precious knowledge, it was going to capture it and make it easier for people to find it.

Next, came the testing phase.

We used Twitter to find some volunteers to test the new bot and help us move this project to the next phase.

For the next three weeks, we built, tweaked and fixed until the bot began to resemble a real product.

You can’t always control timing

The launch plan was in place. We were going to play this one by the start-up book.

First, we needed a pre-launch base. The group of early adopters and beta testers would give us a way to work out any additional kinks and start building hype around the product.

Then, once all the technical issues were mostly behind us we’d launch on Product Hunt early in the morning between Tuesday and Thursday because that’s when you reach the largest audience.

The official launch was going to happen sometime in late March or early April 2016.

You know what they say about the best laid plans.

Luis was still in California when he woke up to an important message on a Sunday morning.

Paperbot had been featured on Product Hunt.

We were still a few weeks away from the planned launch date.

The product was half finished. The landing page didn’t even have an “add to Slack” button.

This wasn’t how it was supposed to happen.

In situations like these, you are faced with a hard choice. Do you let opportunity pass you by or do you dive in and do your best to make it work?

We dove.

A new landing page went up with a link to the install page.

Then, it was time to watch the logs.

For six hours, we fixed bugs, wrote queries, sorted issues through social media and worked out the kinks.

It was a real life make or break moment.

Despite the issues, the early user base was excited. People will forgive a lot of problems as long as you do your best to fix them and communicate clearly and effectively.

A lot of planning goes into building and launching products. We create marketing plans, build user bases and try to account for every eventuality. Those things are important but sometimes they can hold you back.

This forced early launch confirmed our commitment to the community and to each other.

Being in the startup world requires flexibility.

Plans rarely go smoothly. That’s why you need a team that can rise to a challenge and roll with the punches.

Plus, this challenge had a transformative effect on the bot itself.

If we’d waited, we would have lost momentum. Instead, we got a chance to observe how the community used Paperbot and allowed us to add more and broader functionalities outside of a vacuum.

We saw a lot of different use cases that we were not expecting, including collecting RSS feeds into Slack and transforming them into a newsletter and automating the company’s news alert system.

Some teams even asked for API access so they could build on top of our data.

Without the early launch, none of this would have been possible.

Beyond the comfort zone

In life and business, things rarely go the way we plan. All the analysis and research in the world can’t be substitute for the need for timely action and a bit of ingenuity.

Before this launch, we had no bot experience and we always approached product design through a user interface.

Now, we know there’s more to it than that.

When you are building a chatbot, the product goes deeper.

It’s not about looks- it’s about personality.

Your bot has got to be friendly. It needs to keep up with the conversation. It absolutely has to add value to the experience.

We found that users said thank you to Paperbot after a good interaction and that they liked it when it responded in kind.

We learned that charisma and humanity were a key part of functionality.

An effective chatbot had to be a little human, and the early launch helped Paperbot get there.

You can waste a lot of time tweaking and “fixing” a product before releasing it to a product base. Half the time this is just fear trying to dress itself up as something else.

To really take a product to the next level, you need a user base. You need people who are going to mess with, and experiment with, what you are building.

Your users will show you what’s working, what isn’t and what performs well outside of the product creation bubble.

It takes theory and learning into the real world.

Developing Paperbot after the launch

Paperbot continues to grow and change based on the feedback from our user base.

What started off as a way to collect links into an email digest has evolved.

The Paperbot apps, available for both iOS and Android, and the desktop version take its functionality to the next level and align it with Skore’s overall vision.

You can see your team’s link feed, neatly arranged by date and read an optimized version of each article.

The recommended section lets you scroll through the articles that initiated the most conversation between team members.

Every day, we waste a lot of time searching for information, scrolling for that article you can’t quite remember the name of or looking for that really important link your colleague sent you.

Paperbot has a powerful search functionality that lets you search for articles based on words, hashtags or the people who shared them.

Over the last few months, Paperbot has grown into the missing link we were looking for and a lot of that is thanks to community feedback.

This smart bot is helping teams around the world capture valuable information and improve knowledge flow.

It’s a vital step toward ending knowledge waste within organizations and making access to vital information easy and intuitive.

Launching a few weeks earlier than we’d planned was challenging but that challenge helped us create a better, stronger chat bot.

And, it’s taught us a lot about the potential of chatbots and the role they play in creating a powerful, positive user experience.

Working on a bot yourself? What’s your experience like so far? What are some of the most challenging parts of your project?

Sign up for our newsletter and let’s keep the conversation going.

A version of this post was originally published on Chatbots Magazine.