Unilever uses Artificial Intelligence to Recruit and Train Thousands of Employees in Amazing Ways

It is difficult to live in the developed world and not use a Unilever product. Over 400 consumer products brands are manufactured and distributed by the multinational, which includes food and beverages, household cleaning products and personal hygiene.

Artificial intelligence is rapidly becoming a necessity for large organizations with so many processes to coordinate. This includes both research and development, as well as the support infrastructure required for businesses with over 170,000 employees.

It recently announced that machine learning algorithms could detect if you have body odors and can sniff your armpit. Although it may sound like cracking a walnut with a hammer, the technology that has been developed could be used to monitor freshness and help to eliminate food waste .

These smart, public-facing initiatives are supplemented by Artificial Intelligence which is used behind the scenes to screen and assess more than a million applicants for Unilever jobs each year. They have AI-powered tools that will help them adapt to their new role and get the job done if they are selected for the thousands of jobs available.


AI-enhanced recruitment

Unilever employs over 30,000 people per year and processes approximately 1.8 million job application.

This requires a lot of resources and time. Multinational brands operate in over 190 countries. This means that applicants can be found all over the globe. Unilever cannot afford to ignore talent because it’s buried in a pile of CVs.

Unilever and Pymetrics, an AI recruitment specialist, created an online platform that allows candidates to be assessed at home, on their computer, or from their mobile phones.

They are first asked to play a variety of games that will test their logic, aptitude, reasoning, and willingness to take on risk. Then, machine learning algorithms are used to evaluate their suitability for the role they have applied for by comparing their profiles with those of previous successful employees.

Submitting a video interview is the second stage. The assessor is not an actual human being, but a machine-learning algorithm. The algorithm analyzes videos of candidates answering questions for approximately 30 minutes and determines who is most likely to be a good match.

Leena Nair (Unilever’s chief HR officer) told me that the automated screening system had cut down on the number of hours spent interviewing and assessing candidates.

She stated, “We are looking for people who have a sense of purpose. This includes systemic thinking, resilience and business acumen. The games and video interview, based on this profile, are programmed to search for cues that help us understand who will be a good fit at Unilever.

She refers to their video interview analytics for their future leader program and says: “Every screenshot gives me many data points about a person, so we work together with a number partners and use lots of proprietary technology with them, and then we select 3,500 people to go through our discovery center.” Unilever selects around 800 candidates who will be offered a job after spending a day with real leaders.

All applicants will receive feedback, even those not successful.

Nair says, “What I love about the process it that every single person who applies to us receives feedback.”

“Normally, when someone sends a job application to a large corporation it can get into a black hole’. We thank you for your resume and we’ll get back with you. You may never hear from them again.

“All applicants receive feedback on how they did in video interviews and in the game. We also give them feedback on what they need to do in future applications.

“It’s an instance of artificial intelligence that allows us to be more human.”

Unilever may not be ready to give the entire recruitment process to machines yet, but it has demonstrated that it can help with preliminary screening of applicants.


Robots to assist you in your job

After making the grade, another machine-learning-driven initiative is helping new employees get started in their new roles – adapting to the day-to-day routines as well as the corporate culture at the business.

Unabot (natural language processing) bot is built on Microsoft’s Bot framework. It is designed to understand employees and retrieve information when they ask.

Nair says to me, “We joke about it being a man and a woman – Unabot.”

Unabot does not only answer HR questions. Questions about any topic that affects employees should also be answered by Unabot. It is now the front-face for all employee questions – they may ask about their allowances or IT systems – so we learn about what employees care about in real-time.

Unabot learned how to interact with employees and answer questions like where parking is available, when shuttle buses run, and when annual salary reviews will take place.

Unabot cannot filter or apply information based upon who it is talking to, unlike Alexa and other consumer-facing chatbots. It can distinguish between the user’s location and their seniority within the company by the information it sends.

Unabot was initially launched for employees in the Philippines. It is now available in 36 other countries. It was selected to be the next AI initiative in Unilever’s global 190 markets.

Nair says, “It’s an entirely new way of working.” Nair also said, “We never go into a country and say it’s perfect, so let’s spread it in all countries’. We learn from one country and then roll it out to the next.”

All data is currently sourced from within the company, including schedules, policies documents, and employee questions. This could expand to include external data, such as learning materials.

Even though it’s still early, initial analysis shows that the initiative is very popular among staff. 36% of the people in the areas where it is deployed have used it at least once and around 80% continue to use it.

Early on, I learned the importance of providing an easy experience.

Nair states, “So we have learned that you need to make everything that interacts with consumers or employees effortless.”

People interact differently – a policy document is written in one way. It contains three to four pages that outline what employees should do. An employee will often ask simple questions like “How does this affect my life? Where can I find it, and what can I do?”

NLP, machine learning, and especially NLP, can overcome this because it is able to recognize which questions are being asked repeatedly, regardless of how they are presented, and then present the correct information.

HEY! Could we ask you for a favor? Would you share this article with your friends? It costs you nothing and it takes just a second, but means the world to us. Thanks a lot!