⏱ Estimated reading time: 9 minutes

Table of Contents

No managers at Alan… well yes but not what you’re thinking

By Benoït Prioux, Softare Staff Engineer @ Alan

Alan has a written-first company culture. Everything gets written down and work is asynchronous by default. This means no meetings! Everything is public on Slack and Notion, including compensation per level and company strategy. Each one of the 700 current employees has extended ownership and autonomy at work and can organize themselves in the way that suits them best. Employees are all encouraged to provide feedback all the time.

Because employees are owners of the subjects they work on, the responsibility in the organization is highly distributed, clear and communicated. Eployees are encouraged to ask for help on topics they lack expertise on and when they are blocked.

Multiple terms have been developed to designate organizational structures internally:

  • A Crew is a multidisciplinary team of about six people working on the same product;
  • An Area is a collection of about four Crews;
  • Crew Leads follow their Crew day to day operationally;
  • Coaches are experienced people that help employees through their careers and for HR matters.

In the Engineering organization, there has been no Coach despite it growing to 125 employees because the role required a strong experience of the company culture and that employee growth would have caused Coaches to be overwhelmed. Deputies were introduced to reduce the load by having both a talent management and a technical role.

Because this role is demanding, communication is still essentially written, tooling is data-based and connected to knowledge bases to simplify decision making, meeting transcripts are automated are reviews (including self-reviews) are frequent. Basically Deputies experiment with anything that can improve productivity, essentially improving output quality for less effort.

All managers help their team members grow, deliver and fulfill themselves. Some of them were former Individual Contributors who wished to do more for the organization.

Questions and Answers

How did written culture hold up with the advent of LLMs?

Communication has been the standard and always helped structure decisions. Internally, LLMs see use to simplify research tasks. Because it is easy to hold AI wrong, good usage practices are shared internally.

How do employees know each other in this async culture?

Daily meetings are written, but 1 on 1 meetings are done face to face. Plus every six weeks, events are organized in physical locations for employees to engage with each other. To encourage working through complex subjects together, mob programming sessions where features are delivered adter the session are organized weekly in some teams.

Mind the gap: Hours, culture and communication

By Amélie Benoit, Tech Lead Manager @ Busbud

Organizations like Busbud with international reach tend to grow by acquiring solutions that cater to the specifics of local markets ad bring their innovation globally. These cycles of acquisition and integration lead to forming internationally diverse teams, but without issues!

Geographically spread teams often face synchronization and cultural issues. While synchronization issues might somewhat be alleviated by an asynchronous and writing-centric organization culture, culture clash is inevitable and will lead to situations:

  • A Brazilian employee tended to send “kkk” a lot in the internal chats. US employees found this weird and inappropriate with regards to the context. In actuality, for certain cultures “kkk” is an equivalent of “lol”. The employee was communicating that the situation was funny for them but the shorthand, which imitates a laugh, was not understood in the same way by their US counterparts;
  • Canadian employees are polite, considerate and will tend to thank their colleagues for their work. When a Brazilian employee received “thanks for the good work :)”, they were perplexed by the presence of the emoji and though that the message was sarcastic as they are not accustomed to getting thanked for doing their job;
  • French people tend to have a natural tendency to be… frank and direct. Using strong words and expressions like “this code is sh***y” reads as needlessly offensive and confrontational to other cultures;
  • In US style management, managers tend to hide their point in convoluted and diplomatic speeches that highlight what the employee did right. To find the actual issue, the employee must read between the lines and find what is actually wrong;
  • It happens that quite a number of employees do not speak English nor French, which inevitably makes communication more difficult, especially when they have to share their progress verbally. Live translation tools are still not up to the task and can introduce translation errors because they lack context. The language barrier has been broken by improving the translation tooling and having deputies bridging the cultures. Synchronous events are still tricky, though.

These are situations where information loss caused by writing is actually detrimental to the organization. Writing is by definition low-fidelity and relies on culture and biases to communicate messages. Meeting each other is essential and can lift miscommunications. A good read on this: La Carte des différences culturelles: 8 clés pour travailler à l’international, by Erin Meyer.

Being fully remote does not help with most of the team is located in a wildly different time zone. This is the case here, where a French person works with US, Canadian and Brazililan employees. Because the overlap period is slow slim, employees on one or the other side may be blocked until the other side works again. In the French side of the Atlantic, this could lead to people not disconnecting during the evening or night to continue helping teams across the globe. Asynchronous communication does help, but it requires everyone to work asynchronously otherwise this would lead information gaps and confusion because some people get the information sooner than others.

Another issue with time zones is with multi-level on-call procedures which may trigger at the worst possible time for one culture but not the other. An incident occuring during the evening in a panamerican time zone can wake up a European employee in the middle of the night. Hence the importance of well documented and shared procedures and runbooks.

The remote-first culture also created a new issue: solitude. The low fidelity nature of written and online communication can cause people to feel detached from the work they spend a third of their day on! Meeting up with friends, family, interest groups, communities and hobbies is essential to keep this social contact.

How AI helped me get more productive without doing anything for me

By Antoine Dumas, Founder @ Érable Studio

From experience to experience, you get to meet a large crowd and a wide spectrum of ways to engage with content and learn from it. You become sort of a teacher and sometimes your students show that they don’t have a full hold on their solutions. In these situations, being interested in culture and specifically of the technological kind helps popularizing solutions, getting people on productive faster.

It is quite common to hear about “effectiveness” and AI in close proximity of each other. There is a point to be made about that:

  • Effectiveness is compated to a targeted objective;
  • Efficiency also includes the means to reach the end. It’s not about working harder but working better.

Effectiveness does not imply efficiency, but efficiency can grow into effectiveness, because working better cuts down the overall amount of work. Working less to achieve the same result means you achieve way more in terms of quality and quantity in the same amount of time as before!

In some cases, subsituting humans with plain AI also substitutes simple problems with much larger and intractable ones. Instead of applying AI on a problem like jam on a piece of bread, AI can be used as a sidekick, a reviewer, a strategic decision facilitator. AI can be used in the path to efficiency rather than effectiveness. We often default to technological solutions before deeply understanding the nature of our problems, at the risk of implementing a costly tool that misses the target.

For instance, say you receive a lot of HR requests. You could be tempted to add some ticketing solution, boards, status, workflows and the whole organization behind it. Or you could start by tracing the source of those requests: the communication channels, the nature of the requests. Reducing the number of channels and focusing on the most common requests drastically reduces the amount of work. AI helps cutting down on the categorization, HR research, strategy and writing tasks as they are all time-consuming. As for the story, most of the time the issue stemmed from incomplete, out of date, non-communicated or missing documentation.

What if the solution to effectiveness was to trace problems to their source by repeatedly asking “why”, then iterating on the contributors?