How ML6 became the fastest growing AI company in Belgium with uman.ai

What started as a small team of big data experts quickly became an AI consultancy firm with hyperfast growth.

Published on Dec 17, 2019

blogpost header

630

resources in the first 6 months

705

completed learning hours

95%

staff retention

About ML6

ML6 is a fast growing AI consultancy firm — with offices in Ghent, Amsterdam, Berlin and London. They’re specialized in machine learning using the Google Cloud stack and they’re on a mission to empower leading businesses with intelligent technology. With the rising demand for AI and data expertise, the team grew rapidly from 5 to over 60 consultants.

The field of machine learning and data science is evolving very fast. Therefore ML6 needs to spot new technology trends on a consistent basis and adopt their team’s skill set accordingly in order to respond to the new market needs.

As these AI and Machine learning skills are not readily available in the tight talent market, ML6 had to find ways to build this expertise internally. This is why knowledge sharing and skills development form a crucial part of ML6’s company strategy to maintain their edge over its competition.

Knowledge sharing at it's core

In the early stages of the company, ML6 mainly took a top down approach in the way knowledge was shared. In general it was their CTO, Matthias Feys, who made sure that everyone had access to the right information and was up to date with the latest technology and skills.

Scalability issues

As the team of consultants gew, this top down approach quickly became a bottleneck in sustaining their steep growth trajectory. So they started looking for a way to share knowledge more efficiently and develop the skills of their team faster.

Matthias Feys: “As soon as we reached the mark of around 30 consultants, it became hard to scale our approach to knowledge sharing. Especially when opening offices in other countries and working with distributed teams. We needed a solution to sustain our rapid growth.”

Why they need uman.ai

ML6 didn’t believe that yet another tool would solve their scattered knowledge problem.

For this reason they chose uman.ai, as the solution integrated nicely into their current way of working. The uman.ai plugin enables their team to share knowledge directly from the browser while the AI analytics keep track of the skills developed by the team and automatically creates transparency in knowledge and expertise.

Since going live 6 months ago ML6 has seen a drastic increase in the amount of knowledge being shared. On top of this the team now has complete transparency in their skills for the first time ever. ML6 is now integrating these data insights into their core business and talent processes and building a data driven talent strategy.

These results enabled ML6 to become Google Cloud specialized Machine learning partner and the fastest growing AI company in Belgium, ranking in the Deloitte Fast 50.

This is how they did it.

Sharing knowledge directly from the browser to break social barriers

Before uman.ai, a dedicated “interesting” chat channel was used to share links to relevant content. However most of this content was not relevant to everyone in the company: Machine learning related content didn’t appeal to the data engineers and for some people the information was not relevant at this point in time because they were working on other topics or the content didn’t match their expertise level. This way, a lot of important information was missed and forgotten.

On top of that there existed a social barrier of sharing relevant knowledge, especially for new team members. Often people were scared to waste other’s time, “what if the content I recommend is not relevant to this person or to everyone in the group?”


By using the uman.ai browser plugin, team members can securely save any type of content to the ML6 workspace directly from the browser. The intelligence behind uman.ai analyzes the content and distributes it to the people for whom it is relevant, based on their interest profiles. This greatly reduces the social barrier of sharing, as people only have to mark content as relevant and uman.ai takes care of the rest.

“With uman.ai we’re building self learning teams. Our experts can now share knowledge, knowing it will reach their colleagues whenever they need it.”

- Xander Steenbrugge, head of AI @ML6

blog image/video

Discussions happen on the content itself to break knowledge silos

When someone was searching for task specific information, links to internal documentation and relevant articles were shared via chat or email between colleagues. This way a lot of content related discussions and insights lived isolated in one-on-one conversations, not benefiting the rest of the team.

With uman.ai content related discussions are happening in the browser directly on the content itself; leading to less context switching and transparency in information ownership. Now anyone revisiting this content later, benefits from all the insights shared by the team.

blog image/video

Enriching Google Search results with company knowledge to enable on the job learning

A lot of effort was put into documenting best practices to get people up and running faster. Boilerplate code was kept in centralized repositories on BitBucket and best practices were shared via an elaborate folder structure in Google Drive.

Before uman.ai, these documents were hard to find and many people forgot they existed. When they started a new project or were stuck on a specific task most of them searched Google and Stackoverflow for solutions. This lead to a lot of different approaches and wasted time repeating mistakes.

With uman.ai, these documents now surface automatically when people are looking for related information on Google or Stackoverflow, leading to increased productivity and faster on the job learning.

blog image/video

Bringing together team wide expertise with auto-tagged skill channels

Onboarding of new employees is a crucial part to get people up and running faster. Therefore ML6 had created multiple curriculums with recommended content to help develop the minimal Machine learning and data skills required.

However, moving beyond the basic principles and developing deep expertise to work on expert projects requires an increasing amount of information that becomes too niche and specified. The skills and best practices at the edge of this field of expertise are evolving so rapidly that building in depth curriculums becomes impossible.


uman.ai’s smart auto-tagging feature brings together all relevant content shared by colleagues, and shows the internal experts, related to specific skills in the company. Now people can checkout all skill related information when they start looking into a new framework or tool. This way the invested time made by colleagues to curate information can be leveraged at scale and the team can develop new skills faster.

blog image/video

Transferring knowledge at scale with shareable collections

ML6 introduced a mentoring system where senior employees provide career and development guidance for junior people. The main transfer of knowledge happened in person or via chat where the mentee asked the mentor for specific content when learning new skills.

With uman.ai collections, mentees can turn to the mentor’s read later collections at any time. There they can find and consume all curated content saved by the mentor on his learning journey. This creates a scalable way of transferring knowledge without any interference of the mentor.

blog image/video

blog image/video

Core business and talent processes are driven by real-time skills data and stimulate a data driven talent strategy

ML6 didn’t have a centralized, up-to-date overview of people’s expertise. As the team grew it became unclear who possessed what know-how. Staffing mainly happened based on availability instead of expertise and growth potential.

Two yearly growth and performance reviews were not enough to keep track of how people were developing new skills, which made it harder to deliver appropriate growth and training opportunities to every individual.

This lack of transparency also started to impact business development. With technology changing fast it became hard to know which expertise the team had inhouse and if they were developing new skills fast enough to keep servicing their customers.

Matthias: “We knew we had to start capturing skills data, but we really didn’t like the idea of keeping everything in excel competency sheets as they were a lot of work to maintain and not future proof.”

The AI analytics in the uman.ai platform automatically keep track of who is working on which topics. This data gives real-time insights in personal growth and the skills developed by the team. The business can identify particular skills and knowledge that could be relevant to different projects and clients and talent development happens more personalized than ever. This increases employee retention and speeds up development of new skills.

“At one point we started noticing someone was developing an interest in Quantum Computing and Quantum Machine Learning, skills we didn’t yet have internally. From the data insights we were able to proactively start a conversation and find more relevant projects related to the person’s growth desires.”

- Matthias Feys, CTO @ML6

The results

In the first month, over 130 unique documents were shared and over 75 hours of learning was completed by the team. These results show how uman.ai enhances ML6’s knowledge sharing culture, by enabling efficient sharing, effective on the job learning and real-time measuring.

Today ML6 combines the bottom up approach uman.ai offers with top down support from management. The data insights power their personal growth discussions and training decisions. They are using data driven actions to fully embed continuous learning and knowledge sharing into their day-to-day operations.

Matthias Feys: “It’s important that people want to use the solution. They clearly have to see the benefit for themselves. On the other hand, we as management also have to take initiatives to support and encourage our team to develop and share new knowledge. The AI analytics give us detailed insights into how our team is developing new skills, so we can make better, data driven decisions on where to focus in business development, recruiting and training.”

“It’s important that people want to use the solution. They clearly have to see the benefit for themselves.”

- Matthias Feys, CTO @ML6

blog image/video

source - https://blog.ml6.eu/