what's cooking in ai
That AI is the next big thing since the internet is now a common notion. In interviews done with Belgian CIOs by CIONET in early 2020, more than 80% responded with AI as the topic that interests them most. AI is also on top of marketing agenda from all large IT service providers.
As part of the Partnership Benchmark 2020, we wanted to go beyond the hype to find out what is really happening on the ground. AI is a wide term that encompasses many things. But what are the key areas where we see real projects? What competencies are well developed? Who are the key players on the Belgian market today?
AI has evolved beyond Analytics
A few years ago, most companies who claimed to be working with AI were limited to applying AI for Analytics. They used AI to better decipher patterns and trends within data and make smarter predictions. However, the biggest disruption in AI comes from recent advances made in two particular areas of Machine Learning – Machine Vision and Natural Language Processing. These have the power to transform how we work and even how we live. And these are the areas where we see the most interesting projects taking shape today.
Also, giants like Google, Microsoft and Facebook have developed open source machine learning libraries and platforms that are made available to anyone to build AI based applications, giving a big boost to companies active in this field.
AI startups are leading the way
While all large IT service providers are active in this space and are developing their AI capabilities, our study indicates that it is the small AI focused startups who have the most advanced capabilities and projects. In Belgium, companies like ML6, Radix and Faktion have realized projects at large companies and government organizations. Each of them have between 30-70 machine learning engineers and are growing quickly.
Some project realizations
VDAB uses AI to improve the results of job matching for their candidates reducing the time required to fill in introduction questionnaires from 45 to 10 minutes for each candidate. Brussels Airport uses AI to improve passenger experience through Chatbots and accurate prediction for luggage pick up times.
GSK uses AI to speed up vaccine development by using machine vision capabilities. Using machine vision, GSK is able to ascertain how quickly certain living bacteria are able to reproduce on a petri dish in a much faster and more accurate manner.
The challenge remains progressing beyond the POC
While there are a number of examples of AI in production, we found that most initiatives end up getting stuck at the POC stage. This remains the key challenge within the AI domain. So much so, that a metric to judge successful AI providers is the % of their total projects which make it into Production. We discussed how to overcome this challenge with IT leaders and AI service providers:
1. Engage in the topic with the board/ C-Level. Align the initiative to the business strategy and make sure to get top down sponsorship. Do not treat this as an IT project or a standalone innovation showcase.
2. Think big and transformational. For the POC, chose an area that is core to your business and not something peripheral. A successful POC in a core business area has a better chance of adoption and making it into production.
Some Key Company Profiles
While there are a number of small companies active in this area, we highlight the two most important providers in this space. The other notable companies are: Faktion, Sagacify, Dataroots, Kantify and ReImagine.
Founded in 2013, ML6 started with a focus on Big Data / Analytics. As of end 2017, they decided to focus completely on Machine Learning and also changed their name from Datatonic to ML6. They have close to 70 people and probably the widest client portfolio in terms of the number of industries that they serve. They develop complete machine learning solutions and cover both Machine Vision and NLP. They like to do projects that are core to the business or technologically complex projects. ML6 work only with Google’s AI platform and are a preferred partner. Key clients include Belfius, TVH, Bekaert, Pfizer, Melexis, Fednot, ING and UZ Leuven.
Radix was founded in 2018 with a focus to help clients leverage AI in achieving their business objectives through AI. They have grown to 30 people and focus on both Machine Learning and Machine Vision. They offer complete machine learning solutions but are not limited to any one platform. Some of their marquee projects have received very positive coverage in the press – namely VDAB and GSK. Other clients include Brussels Airport, Atlas Copco and Mediahuis.