Inside Phenom’s Approach to Contextual AI for Talent Managem…

Finding the right job has always depended on timing, skill, and opportunity. Yet traditional HR systems, built on transactional flows, often leave a gap between the talent and the opportunities they deserve. Phenom, a global HR technology company, aims to close that gap through an AI-first approach across the entire talent lifecycle.

The company’s goal is ambitious: help a billion people find the right work. In an exclusive conversation with AIM, Sivanand Akella, senior vice-president (SVP) of engineering at Phenom, shared how the company is rooted in applied AI, not just surface level automation. 

Phenom builds contextual, intelligent systems that provide personalised and predictive experiences, instead of layering automation onto outdated workflows.

From Chatbots to Contextual Intelligence

Phenom’s Intelligent Talent Experience platform blends experience, automation, and intelligence, to support the entire hire-to-retire journey. 

“AI is already delivering measurable impact,” Akella said. He pointed to the platform’s Fit Score, a feature that drastically cuts screening time by ranking candidates on skills, experience, and location. Conversational chatbots and automated scheduling agents take over repetitive tasks, saving thousands of recruiter hours.

But Akella emphasised that automation is just the beginning. True transformation comes from contextual intelligence. 

“HR cannot use ChatGPT for hiring,” he said, adding “It must be contextualised.”

Phenom’s edge, he explained, lies in its ontology-driven infrastructure that understands labour markets, specific job functions, and enterprise workflows. Phenom goes beyond generic automation by pairing this contextual foundation with generative AI. The result is strategic workforce insights and long-term engagement.

Tackling Keywords and Bias in Hiring

One of the most persistent challenges in recruitment is bias introduced by keyword filtering. Akella was clear that Phenom’s approach is more intelligent. “Our AI is deliberately built to move beyond keyword-based filtering, which is one of the biggest contributors to bias and missed opportunities in hiring.”

Instead of rejecting a CV for lacking a specific keyword, Phenom’s AI leverages skills ontologies, embeddings, and contextual reasoning to understand the meaning behind an individual’s experience. 

For example, a “business journalist” applicant might not mention “business” explicitly, but the platform maps their competencies and adjacent roles through graph-based reasoning. “This ensures candidates are evaluated for what they can do and have done, not just what keywords appear in their resume,” Akella said. 

Retrieval-augmented generation (RAG) grounds these assessments in verified skills data and labour market intelligence. The result is a fuller, more accurate view of a candidate’s potential.

Bias detection loops, explainability layers, and recruiter feedback monitoring ensure the system not only highlights qualified candidates but also nudges hiring teams toward fairer, more transparent decisions. By combining ontologies, RAG, and bias-monitoring agents, Phenom ensures that strong talent isn’t overlooked because of surface-level wording.

As one example, Mastercard’s partnership with Phenom helped drive 900% more candidate profiles, 141,000 more leads, and 11% higher application conversions than industry averages. The company also claims that Mastercard saw an 85% faster interview scheduling and drew in over 9,00,000 members in the talent community.

Building a Future-Ready AI Infrastructure

Underneath, Phenom operates a hybrid AI architecture, combining proprietary models with fine-tuned LLMs and open-source frameworks. 

More than 25 specialised AI agents power tasks from candidate sourcing to fraud detection, orchestrated through tools like LangGraph. To keep systems reliable and safe, Akella emphasised, Phenom embeds human-in-the-loop validation, explainability, and observability.

This focus on scalability extends to engineering as well. Kubernetes, Kafka, and event-driven design ensure the platform can handle millions of daily interactions while maintaining speed and resilience. “Scalability is about more than training larger models—it’s about the engineering backbone,” Akella explained.

Looking ahead, the company claims to prioritise retrieval-augmented generation (RAG) and graph-based reasoning to ground AI outputs in structured knowledge. Its talent knowledge graph, spanning over 40 ontologies, enables reasoning over skills, jobs, and career flows. For recruiters and managers, this not only means faster answers, but contextual, defensible recommendations for workforce planning and internal mobility.

While competitors like Beamery, Eightfold, and Workday, all market themselves as being AI-first, there could be differences in the features being offered across the platforms. For instance, Eightfold leads in deep talent matching and predictive talent intelligence, as per some user reviews. Companies might need to review various aspects of these platforms before embracing them.

Human-Centric AI for HR

Despite the sophistication of its technology, Phenom’s philosophy is simple: AI should amplify human potential, not replace it. “AI augments decision-making, but the final word is always human,” Akella stressed. Routine screening, rediscovery, and ranking may be automated, but empathy-driven hiring decisions remain firmly with recruiters and managers.

For Akella, the real shift is redefining HR from an administrative function to a driver of business strategy. With AI as the backbone, metrics move from time-to-fill to long-term retention, engagement, and mobility. In his words, “The future of HR is both intelligent and deeply human.”

The post Inside Phenom’s Approach to Contextual AI for Talent Management appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source