Signal Labs Launches SignalWork, Putting People and AI Agents on the Same Work Ledger

On Tuesday, October 14, Chicago-based Signal Labs announced SignalWork, a work allocation platform built for CHROs, CIOs, CTOs, and CEOs who are facing a new kind of staffing question. It is no longer just how many people the company needs, but which mix of people and AI agents should do each unit of work, and what each choice costs.

The company describes SignalWork as a single ledger on which work is allocated across humans and AI agents, then executed within enterprise-defined controls. Rather than a reporting dashboard, it is designed to rehearse, decide, and act: simulating how a change in staffing or automation plays out before resources are committed, assigning the work, and tracking whether the outcome matched the prediction.

Signal Labs is not a newcomer to workforce data. The company is backed by Lightspeed Venture Partners, and in July 2026 it acquired BrassRing, a veteran cloud-based applicant tracking system with customers in 172 countries. The launch arrives one day after workforce analytics firm ActivTrak opened early access to a Model Context Protocol integration that connects behavioral work data to ChatGPT and Claude, making this week an unusually busy one for the office software stack.

The problem: planning tools built for a people-only workforce

Signal Labs argues that workforce planning today is split across an applicant tracking system, an HRIS, and finance software, none of which talk to each other or offer a predictive view. Two failure patterns follow, the company says. First, AI spend climbs unchecked because token cost is never measured against the outcome it was meant to produce. Second, roles get automated wholesale, carrying away the relationships and judgment a person had built up, which can quietly break the workflows that depended on that person.

"You are running a mix of people and AI agents against goals that used to be staffed by people alone," said Rajeev Ronanki, Founder and CEO of Signal Labs. "You need to understand how that work fits together and what happens when you change who does it."

Signal Labs' answer is what it calls institutional world models: living representations of how an enterprise operates, covering its goals, workflows, workforce capabilities, relationships, costs, and constraints. Executives use them to simulate and stress-test strategic choices inside a virtual replica of the organization, a digital twin for workforce decisions, before executing them for real. Example rehearsals: the effect of a 10 percent price increase on fourth-quarter revenue, whether a new office location would be profitable, or how a 15 percent market drop would hit the business, drawing on internal systems plus external data such as news, weather, and geopolitical events.

What the platform actually does

SignalWork connects internal systems such as BrassRing, Workday, and SAP into a single view of the workforce, then recommends what work should be prioritized and which mix of human judgment and machine capacity should perform it. The launch announcement lists six capabilities: rehearse workforce choices before committing resources, allocate work and attention across people and agents while surfacing decisions that need human judgment, execute within enterprise controls, source third-party AI agents from a marketplace onto the same ledger, manage AI spending against outcomes (token burn weighed against results), and retain a traceable decision record feeding outcomes back into the models (Decision Memory).

The first users will be Signal Labs' existing SignalOS customers, Infinite and VetJobs. Jamie Dale, Senior Director of Technology and Data Infrastructure at VetJobs, said the nonprofit had already seen proof that SignalWork eases grant compliance work by 50 percent, and that prior automation through BrassRing had saved the equivalent of two full-time employees every week in recruiting operations.

"SignalWork is intended to move the focus from time-to-hire toward time-to-outcome," said Ethan Ray, Senior Research Analyst at 451 Research by S&P Global, describing a system that reads the funded work portfolio, identifies units of work called missions, and connects them with human and agentic AI capacity in a unified ledger.

The companies that win the next decade will not be the ones with the most AI agents. They will be the ones that know, down to the task, where each agent's work ends and human judgment begins.

A second front: workforce data moves into the chat window

One day before the SignalWork launch, ActivTrak opened early access to ActivTrak MCP, an integration built on the Model Context Protocol that connects workforce productivity and behavioral work data with AI platforms. It is available to eligible customers on ActivTrak's Workforce Management and Productivity Optimization plans, supports Claude and ChatGPT at launch (Microsoft Copilot and Google Gemini are planned), and costs nothing extra during early access.

The idea is to let leaders investigate work patterns conversationally, alongside business data from CRM, HR, and ticketing systems. A support manager could compare ticket resolution times with application usage to ask whether delays reflect staffing constraints, inefficient processes, or technology problems. Administrators retain control over whether the integration is enabled, which AI platforms are permitted, and how usage is audited.

ActivTrak's own research explains the product bet. Its 2026 State of the Workplace report, based on more than 443 million hours of digital work activity across 1,111 organizations and 163,638 employees, found that 80 percent of employees used AI tools at work, while focus time declined and collaboration and multitasking increased. Adoption, the data suggests, is not proof of better working conditions. Gartner reported in September 2026 that only 22 percent of organizations had successfully scaled AI across multiple business units or adopted an AI-first approach.

What it all means

Read together, the two launches describe the same shift from opposite ends. SignalWork is aimed at the decision: where should people and agents be deployed, and what happens if the mix changes? ActivTrak MCP is aimed at the measurement: what is actually happening in the workday, and is AI changing it for the better? Both assume the unit of analysis in workforce planning is moving from the job to the task, from headcount to a ledger of allocated work.

There are real risks attached to that ledger. Behavioral work data is sensitive even when used for planning rather than performance reviews, and organizations will need to weigh access permissions, data handling, and employee privacy. Activity metrics do not capture every aspect of contribution, and a correlation between app usage and strong performance does not prove causation. AI-generated workforce analysis needs clear data provenance and human review, especially when it affects employees' livelihoods.

Signal Labs itself flagged one of the sharper risks: automating a role wholesale can take the relationships and judgment a person built up, and quietly break the workflows that depended on that person. A ledger that recommends the cheapest mix of humans and agents is only as good as the costs it can see, including the ones that never show up in token counts.

Whether institutional world models can predict organizational outcomes accurately enough to be trusted remains unproven. SignalWork is brand new and its first customers are just getting started. But the direction is unmistakable: after a decade in which HR technology recorded what the workforce was, the race is on to build software that rehearses what it could become, and executes on the answer.