Photo courtesy of Saad Bin Shafiq.
Opinions expressed by Digital Journal contributors are their own.
In November 2025, Foundation Capital published a thesis that went viral in venture capital circles. The firm identified “context graphs” as AI’s trillion-dollar opportunity, describing platforms that capture not just what decisions companies make but why they make them. The concept addressed a gap in enterprise software. Systems track actions but lose the reasoning behind them, and that institutional knowledge disappears when employees leave.
Saad Bin Shafiq had been building exactly that infrastructure since October 2023, eighteen months before investors had language for what he was creating. While venture capital was still debating whether AI would disrupt hiring software, Bin Shafiq was already deploying a system at Fortune 500 scale that learned what success meant inside individual companies and compounded that knowledge over time.
The category now has a name: The Talent Intelligence Layer. It represents a fundamental shift in how enterprise software approaches people’s decisions, and Bin Shafiq’s NODES platform defines what that category looks like in practice.
What talent intelligence actually means
The HR technology market will reach $41 billion by 2035, but nearly all of that spending goes to software built around workflows rather than intelligence. Applicant tracking systems move candidates through hiring stages. HRIS platforms store employee records and track performance metrics. Talent management suites coordinate both but don’t fundamentally change what either system does.
None of these tools answer the question that determines whether hiring succeeds or fails. What distinguishes people who thrive at your company from people who don’t? Traditional HR software tracks who applied, who got hired, and how they performed. It cannot identify patterns that predict success because it was never designed to learn from outcomes.
“Every ATS can tell you who applied,” Bin Shafiq explains. “Every HRIS can tell you who got hired. None of them can tell you why your best people succeed nor find more people like them. That’s what we built. The layer that captures the ‘why.’”
That distinction defines the category. Workflow software manages processes. Intelligence software learns what works and gets better at predicting what will work next. The difference shows up in results. NODES processed over 710,000 candidate applications for CNO Financial Group, which hired more than 4,000 people. The platform predicted with 80 percent accuracy which of those hires would become top performers, validated over 10 months.
Traditional ATS platforms cannot make that claim because they were never built to test their own effectiveness. They track applications and hiring decisions but have no mechanism to learn whether those decisions were correct.
Building before the market had language
Bin Shafiq started NODES in October 2023 without the vocabulary that venture capital would develop eighteen months later. He saw the problem from lived experience. After applying to 700 jobs and receiving one acceptance, he understood that hiring systems filter candidates based on pattern matching to resumes rather than actual capability.
“I am the person their algorithms would filter out,” he says. “699 times, I watched systems reject me not because I couldn’t do the work, but because my resume didn’t match the pattern they were trained to recognize.”
That insight became the foundation for a different approach. Instead of matching keywords, NODES analyzes patterns in companies’ existing top performers to understand what success actually looks like at that specific organization. The system then evaluates candidates against those learned patterns rather than generic job requirements.
Building this required technical architecture the HR tech industry had not developed. Bin Shafiq designed 78 specialized AI agents that coordinate across CRM systems, HRIS platforms, and applicant tracking software. Each agent handles distinct analytical tasks while operating entirely inside customer infrastructure. The platform runs on open-source models customers own and learns continuously from outcomes without sending data externally.
The multi-agent architecture enables specialization that single systems cannot achieve. Individual agents focus on pattern recognition in top performer data, bias detection, skills inference, trajectory analysis, and outcome prediction. Each operates independently but shares outputs through a coordination framework, allowing the system to handle evaluations that would overwhelm monolithic models.
Why category timing matters
Bin Shafiq’s two-year head start matters because talent intelligence creates compounding advantages. Every hiring decision generates outcome data. Six months after each hire, NODES analyzes whether its prediction was accurate. Correct predictions strengthen the patterns that generated them. Incorrect predictions adjust the model. Each deployment adds patterns. Each quarter validates or refines them.
A competitor entering the category today would need years of Fortune 500 deployments to build equivalent pattern libraries. By then, NODES will have years more data. The gap widens instead of closing, which is characteristic of infrastructure that learns rather than infrastructure that simply executes.
“Once you understand what makes someone successful in a senior role, you can identify which people in mid-level roles have the highest probability of succeeding if promoted and what specific skill gaps they need to close,” Bin Shafiq explains.
This reveals why talent intelligence represents a category rather than a feature. Traditional HR software could add AI capabilities without becoming intelligence platforms. They would still track workflows and manage processes. Talent intelligence requires fundamentally different architecture because the system must learn from outcomes, compound knowledge over time, and improve predictions continuously.
Market validation for a new category
CNO Financial Group provided the proof that talent intelligence works at enterprise scale. The Fortune 500 insurance company with roughly $4 billion in annual revenue had rejected six AI hiring vendors over 18 months because each required sending data outside company infrastructure. NODES integrated with their existing Avature recruiting platform in under a week. CNO’s legal team approved deployment in 17 days. The company rolled out NODES across all 215 locations without pilot testing.
Results validated the category thesis. Hiring timelines dropped from 127 days to 38 days. CNO documented $1.58 million in savings over three quarters. The Director of Field Sourcing told Bin Shafiq, “We’re screening thousands in hours instead of weeks, and the quality is measurably better.”
More significantly, the platform’s 80 percent accuracy predicting top performers demonstrated that learned patterns outperform keyword matching and resume screening. Traditional hiring tools cannot measure their own accuracy because they have no mechanism to validate predictions against outcomes.
That validation drove inbound interest from over 20 Fortune 500 companies in banking, insurance, and defense. All face constraints traditional HR software was not designed to handle. They need systems that operate inside their infrastructure, learn what success means at their specific organization, and improve continuously without external data transmission.
The enterprise AI adoption rate reached 82 percent in 2025, but adoption in regulated industries lags because most AI tools require compliance approvals that take months. Bin Shafiq built for that constraint from the beginning, which explains why NODES defined the category while larger companies were still debating whether the opportunity existed.
