Searching for an employment prescreen tool that enables hiring managers to do more with less. The tool is required to prescreen more candidates per position while gaining deeper insights to effectively rank each applicant to expose stronger talent.
Green Fields Solutions, Inc. is a technology solutions firm focusing on developing artificially intelligent applications. Headquartered in Mesa, AZ, Green Fields Solutions seeks to offer easy to use tech products. CEO Ryan Wilson started the firm to give a platform for tech developers to hone their skills and network with like-minded programmers. The current Green Fields Solutions team of tech developers and creative designers have a passion for making technology accessible for everyone.
Human resource departments and recruiters are inundated by resume submission. Research shows that each resume only receives 4 to 6 seconds to catch the attention of a hiring manager. How is it possible to give every applicant an opportunity to express the reasons why they are the best candidate? Or better yet, how do hiring managers expose qualified candidates with the right mix of skills, experiences, and personality from the crowd?
The goal is to create an effective conversational AI application that enables HR departments to prescreen more job applicants while performing a more in-depth analysis of each candidate. The conversational AI application will analyze each candidate's language to determine relevant keywords, correctness in answer, and personality insights.
Designed and developed an artificial intelligence application that is advancing candidate assessment. The application expands analysis opportunities while saving time performing job applicant evaluations. The solution provides deeper analysis in the relevance of candidate answers, higher levels of personality matches, and advanced machine learning data to build smart HR departments for the future.
* Solution strategist
* UX designer
* Conversational designer
* Graphic design
Learning User Profiles
Creating personas and journey maps help design for people besides ourselves. We can also use journey maps to learn more about how users will flow through 'happy path' funnels.
Overview Solution Architecture
Call Flow Diagram
Drawing diagrams of the conversational call flow provide stakeholders a 30,000-foot view of the plan. This enables team members a chance to comprehend the plan. Stakeholders have the chance to confirm technical feasibility, wordsmith messaging, and the team's overall understanding of the business goals.
IBM Watson Assistant
I decided to create the bot using IBM Watson Assistant. The primary reason for choosing Assistant was because of the control over the
dialog nodes. The bot's main objective is to simply transcribe and store the candidate's answers from the primary screening questions.
The candidate answers will be concatenated and later analyzed by IBM Watson Personality Insights. Initially, the question and answer
will allow for caller digression. However, the conversational aspect of the bot will activate near the end of the Q&A dialog where a
candidate might engage with additional questions about the company or specific to the position.
Natural Language Processing
Focusing on bot effectiveness we create a confusion matrix to determine where overlapping is occurring and correct it. The fine-tuning intent performance will come from training phrase audits and revised labeling. It also comes from auditing and fine-tuning entities, while caring over the same diligence in the annotations.
Annotating entities within intents
Intent Classification Modeling - Confusion Matrix
Training phrase - utterance labeling
IBM Watson Personality Insights
IBM Watson Personality Insights is the final element and is a tool that analyzes written text and produces insights relating to the writer's personality. The Big Five personality characteristics represent the most widely used model for generally describing how a person engages with the world. A recruiting company creates a ground truth baseline of all the team members' personalities and compares the data to incoming candidates to assess matching variances.