Inside Ai Part I: Starting from First Principles

How we reinvented Automated Insights as a SaaS product company by embracing Agile and going Org-less and Manager-less

2016 was a big year at Automated Insights. We launched the first SaaS-based Natural Language Generation platform (check it out: Wordsmith), restructured how we are organized, got rid of the role of “manager”, implemented a new company-wide mentorship program, and overhauled our performance review, compensation review and bonus processes.

Whew! It was a lot for the company to absorb in a short amount of time (the first of many lessons learned), but we are hitting our stride in 2017 with several months of accelerating revenue growth. With almost a year of these changes under our belt, it is a good time to reflect and talk about what worked and what didn’t.

Selfishly, posting this to Medium provides a convenient location I can point to when I’m asked about Ai’s organizational approach, which can take a couple hours to fully explain because of its departure from the norm.

So what was the impetus behind the changes? As we transitioned to a product-focused company with Wordsmith versus a services-focused company before Wordsmith, the time was right to revamp several aspects of how we operated. I started to see some issues creep into the company that plague most growing startups, and I wanted to address them early.

I worked at large companies for the first 15 years of my career, and much of what we do at Automated Insights is a direct result of trying NOT to stand on the shoulders of those giants.

If anyone tells me we should do something because “that’s how other companies do it,” I’m instantly skeptical. A lot of traditional company culture, even what is commonplace with startups, originated from larger companies trying to scale or deal with their massive size. Most of it doesn’t work well for smaller companies — actually, I don’t think it works well at big companies either, but definitely not for sub-100 employee companies.

Inside any thousand-plus person company there is so much inefficiency, redundancy, and focus on minutia that builds up over time, I believe these companies could randomly layoff 10% of their workforce and not notice an appreciable impact to their business. If they could get away with it from a morale perspective, periodic random layoffs at big companies would be a good way to combat the inevitable employee-creep that happens over time. Some companies do a version of this as their growth starts to plateau.

I saw big company inefficiency first hand early in my career. It was evident in the uncontrolled expansion of middle management. Or employees whose primary job was to schedule meetings or take notes at meetings or talk about what meetings to schedule next. Then there were people who simply reported on the status of things. Everyone knows how slow progress is inside a big company, which provides an easy excuse why it is takes so long to get anything done. Let’s have a meeting next week to check on status!

Inefficiency has insidious ways of protecting itself and spreading.

Given my inherent skepticism of how most companies operate, I decided early on with Automated Insights to start from first principles when it came to running the company. Instead of replicating traditional company practices, we’ve taken our own path and experimented with different approaches. Not everything we do is untraditional, but we are thoughtful about what we implement and don’t take the easy road just because “that’s how everyone does it.”

I’m writing a series of posts that describe both the what and why of the changes we made at Automated Insights over the past year as we transitioned from a services company to a product company.

Part II looks at how we introduced Agile into our engineering organization using Spotify’s Squad model.

Part III will go into detail on how we remapped our organizational structure to better align with the customer journey.

Part IV will cover how we got rid of the traditional “manager” role and replaced it with two new roles.

Part V will cover some of the lessons learned over the past year implementing these changes as well as other areas that we still haven’t figured out (quick preview: I not a fan of the concept of employee titles).

If you like this article, please let me know by clicking the ❤️ button below!