AI in Workday Peakon: The emerge of Engagement Avengers

Relying on my own (limited) perceptivity and the traditional tools made the process incredibly challenging. Moreover, it took me weeks before I was able to deliver final reports that, by the time they were ready, were already somewhat outdated.

Well, that was just the tip of the iceberg! Ordering the translations was like a global game of hide-and-seek. All in all you'd sweat and pray that ‘employee satisfaction’ didn't turn into something completely different.

Luckily the times have changed. Today with just two clicks I can switch on Peakon, an employee engagement software, platform, and methodology. Which uses heavily-researched scientific approach to measure employee sentiment, creates ‘fancy’ analytics, and provides actionable feedback.

If you too are interested in Peakon, you need to know, however, that it is helpful only for those organizations that are convinced that ‘happy’ employee is a good employee. In other words, with Peakon, you should expect to receive negative feedback. Only then you can actually engage with your employees and address their concerns. As a result, you might, in fact, be able to implement both quick fixes and long-term strategies.

NLP the super-smart librarian

Before jumping into concrete examples of AI in Peakon, let’s address the Peakon underlying AI technology - Natural Language Processing (NLP).

Imagine NLP as a super-smart librarian who can read and understand every book in the library, no matter the language. First, our librarian uses magic scissors to cut up sentences into smaller pieces, called tokens. These tokens are like individual words or phrases, turning a sentence into a bunch of word confetti.

Next, the librarian puts on her detective hat and tries to figure out how these tokens relate to each other. She looks at the context and tries to understand the meaning and intent behind the sentence.

To do all this, the librarian joins her forces (knowledge of languages) with her super-smart IT friends (machine learning and deep learning models). Together, they use huge amounts of data to become even smarter and better at understanding languages.

AI in Workday Peakon: The Engagement Avengers

Maybe not as entertaining as Marvel or Disney superheroes, also Peakon has developed its own warriors. Let’s call them the Engagement Avengers - for their ability to understand and utilize sentiment data, for the good of both employers and employees.

The Fortune Teller (Attrition Prediction) foresees who might leave the company next and helps you keep your top talent. The Speed-Reader (Comment Summarization) cruises through feedback in the blink of an eye. The Polyglot (Comment Translation): turns any written character into something you can understand. No message can be hidden from the Word Inspector (Semantic Search). The Accountant (Semantic Topics) organizes comments into clear categories, making sense of the chaos. The Personal Trainer (Strengths and Focus Areas) highlights strengths and areas for improvement. And finally, the Scorekeeper (True Benchmarks) shows where you stand in the global league of employee engagement.

• The Fortune Teller (Attrition Prediction) uses complex formulas trained on millions of survey responses. He searches for clues in 5 key areas, so he can figure out who might be planning to resign. He looks into job satisfaction, promotion history, work environment, length of service, and performance ratings. Then he checks the risks of leaving for different groups in the company, like departments or teams. Finally, he compares these group risks to the company's overall risk.
  
  
• The Speed-Reader (Comment Summarization), who was trained by the Great Wizard (Google Vertex AI model), uses a secret spell (Generative AI) to create summaries of large volumes of comments and magics up main themes. She does it after each survey for certain groups: Detractors (scores 0-6), Passives (scores 7-8), and Promoters (scores 9-10).  If you are, however, on good terms with her (manager with access rights), she can provide you a summary on demand.
  
  
• The Polyglot (Comment Translation) obtained a great power of AI Translations. His main task is to make sure no comment is lost in translation by displaying translations automatically as a default. For those, however, who prefer a hands-on approach, they can ask the Polyglot for a special translation button that appears below each foreign comment. But how this magic happens? The Polyglot sends comments to Google's Cloud Translation - Advanced service, using the Neural Machine Translation (NMT) model. These ways the Polyglot continues to bridge language gaps, ensuring every voice, in the land of Workday Peakon, is heard and understood.
  
  
• When users need to find relevant comments, it is the moment when the Word Inspector (Semantic Search) jumps into action. The Word Inspector uses a special tool to turn comments and search queries into unique codes. These codes are stored in a database. When someone searches for something, The Word Inspector looks through the database using these codes. Like a detective, The Word Inspector sorts the results by how closely they match the search, from most to least relevant.
  
  
• When the land of Peakon seems to plunge into chaos, we call for the Accountant (Semantic Topics). This superhero The Accountant uses advanced language models to sort comments into specific Workday topics .The Accountant’s superpower is hidden in multiple NLP models, that help to classify each comment into one or more Workday-defined topic classes. These models include: keyword search queries, semantic search queries, keyword-based topics in database, semantic topics bank, best practices Google Search terms leading to Workday, use cases of various products inside Workday, customer inputs, and employee voice questions.
  
  
• The Personal Trainer’s (Strengths and Focus Areas) life mission is to help leaders identify areas for improvement (focus areas), as well as underline success stories (well-performing drivers). After each survey round, the Personal Trainer provides managers’ dashboards with suggestions. They accept or reject these suggestions. When accepted, the focus areas are visible to employees on their personal dashboards. Above all, Personal Trainer’s force comes from a model trained using data coming from only from one specific customer.
  
  
• Last but not least, the Scorekeeper (True Benchmarks) makes sure that the company does not live in own bubble. He uses global regression models to collect and analyse driver scoring data across all customers, adjusting for company size, company sector, employee tenure, age, department, level, office location, and gender. The Scorekeeper's capabilities stem from a blend of offsets and weights that reflect typical answering patterns and attribute distribution. These were generated for every question set and driver, as well as for each standard segment. The models were developed using mixed-effects linear regression, based on data from the last 12 months.

Saving the World of Employee Engagement

Gone are the days of fighting with spreadsheets and outdated reports. With Peakon, organizations can measure employee sentiment, create insightful analytics, and provide actionable feedback with just a few clicks. The underlying AI technology, Natural Language Processing (NLP), acts like a genius librarian that understands and organizes vast amounts of data. Peakon itself has a constantly growing team of ‘superheroes’. ‘The Engagement Avengers’ transform the chaotic world of employee engagement data into a world of clarity and understanding. These AI-powered heroes ensure that employee voice is heard, and their concerns can be addressed.

Sit back, relax, and let the Engagement Avengers save the day in the land of Workday Peakon. My team at Tietoevry Tech Services, can help you to achieve exceptional employee experience thanks to state-of-the-art tools. Reach out via the contact form or check this page for more information.

Workday community materials and Heena Ruparel contributed to this blog creation.