HR Departments face multiple challenges like responsibility for all phases of employees, chronic understaffing, cross-departmental requirements, and time pressure - Using AI can be a solution. This is how Taledo's artificial intelligence supports recruiters in their daily business.
Convincing candidates in a personal interview to change jobs and possibly relocate will be difficult to automate in the foreseeable future. For specific tasks however, intelligent systems can save a lot of time and resources in candidate selection and in determining the fit between the job profile and the candidate's qualifications.
Also for these specific tasks, challenges are manifold. There are many ambiguities, e.g. the meaning of a job title varies greatly between small and large companies. Experienced candidates no longer mention skills they take for granted, whereas junior candidates tend to sell even basic knowledge as expertise. In addition, valuable information is often not explicitly stated, but has to be inferred. For example, a developer in FinTech has probably had experience with high workloads and scaling, but might only show features and technologies in the CV.
Even big platforms like LinkedIn are not perfect. If one searches for a keyword such as "Java" without any further restriction, the entire profile is searched and recruiters who state "search Java developer" in their profile also appear. Furthermore, LinkedIn does not impose any restrictions on the input, so any spelling of skills can be added - even imaginary word constructs are allowed.
There are various approaches for matching candidates and job offers. Classically, there is the syntactic search, in which exact word wording is matched. The simplicity of this approach makes it still being used today. Its disadvantage lies in the strictness, as it can easily miss matching results.To soften filter criteria, one can consider word similarities. For example, synonym dictionaries can extend the search to related terms, and normalization methods such as Stemming or Lemmatization can trace words back to their root form.
Further flexibility is offered by Boolean searches, which allow logical expressions such as "Java or Spring". Phonetic similarity and word distances can be considered to deal with typos.
Semantic searches, on the other hand, are more advanced, where the goal is to understand the content of the search input. A classic representative of this approach are ontologies and knowledge graphs, which establish relations between words. Furthermore, there are embeddings, the mapping of words or texts onto a vector space, so that distance calculations can be used to search for similarities. For a detailed understanding of your own domain, it is important to proceed iteratively and exploratively.
Taledo, EU-funded for its Artificial Intelligence, uses semantic search through NLP methods and its own algorithms.
At the satellite event of the Berlin Science Week, the keynote of AI Monday focuses on Talent Management: How AI has an impact on companies trying to manage change of the talent they need to retain and to recruit. Marcel Poelker introduced some details of their technical solution specifically focusing on the parts supported by AI technology.
Taledo's AI is using semantic search, which aims to understand the content of the search input.
Using NLP methods and onthologies, Taledo's matching is trained on the interview invitation and provides scoring to find the best fit.
The EU-awarded matching algorithm in combination with human recruitment experts creates the perfect solution.
Durch das hybride Modell vereint Taledo moderne Technologie und menschliche Intuition, Erfahrung und Emotionen. Unternehmen profitieren von hoher Effizienz durch geringerer “Time-To-Hires” sowie “Cost-Per-Hires”.Read more
Taledo-Mitgründer Marcel Poelker sagt: "Layout wie Grafiken und Farben oder auch Schriftgröße und - art sind für das Matching in den meisten Fällen irrelevant - hinterlassen aber einen bleibenden Eindruck beim menschlichen Recruiter."Read more
In den USA schon etabliert, nehmen Computer-Softwares auch in Deutschland eine immer wichtigere Rolle im Bewerbungsprozess ein. Kandidaten müssen dann sowohl die KI als auch später den Personaler beeindrucken.Read more
Recruiting-Abteilungen kämpfen nicht nur mit chronischem Zeitmangel, sondern es steht auch häufig der Vorwurf der Diskriminierung im Raum. Inwieweit können Algorithmen helfen?Read more