Findem launches Studio — AI agents that return finished people work

Findem launched Findem Studio, combining its people-intelligence graph with expert-backed agents that deliver finished work such as succession plans, market maps, hiring briefs, and skills analyses for teams to review.

SOFTWARE desk — enterprise AI is shifting from search answers to finished work with evidence. People decisions are a high-stakes test of that claim.

What the post says Studio returns. It is meant to turn a complex people question into finished work a team can review, defend, and act on. Studio combines Findem’s own people intelligence with expert-backed AI agents. The work named on the page is a succession plan, a market map, a leadership benchmark, a hiring brief, a skills gap analysis, and an org chart. A succession plan is a written path for who could fill a key job. A market map is a picture of who works where in a field. A hiring brief is the write-up of a role before a search starts. A skills gap analysis compares the abilities a team has with the abilities it needs. An org chart is the diagram of who reports to whom. People intelligence, on this page, means structured facts about people, companies, and time. An agent, here, is software that produces a piece of work, not only a chat answer. The post says this is a stronger starting point than creating and checking that work from scratch. File the list as the company’s. This desk did not open a plan.

Three requirements, still the page. Each Studio agent is supposed to bring three things together. The first is context from Findem’s expert-labeled people data. Expert-labeled means a person who knows the field tagged the fact, not only a model. The second is a proven method from a leading practitioner. A method, here, is the written way the work is supposed to be done. The third is a set of checks that connect a conclusion to the evidence behind it. The example on the page is a succession plan for a critical role. The post says the agent does not stop at a candidate list. It judges possible successors against criteria an expert defined, and it writes a plan that explains the recommendations. A team can look at the evidence, challenge the conclusions, and change the plan before anyone acts. File the example as the company’s. This desk did not request one.

Three ways to use it, in the page’s order. First, run a ready agent. The page says you pick an agent shaped by a leading practitioner’s method and you receive finished work with the evidence attached. The examples it names are succession plans, candidate intake documents, hiring briefs, benchmarks for a chief human resources officer and for a vice president of talent acquisition, market maps, and talent intelligence reports. The page shortens the first title to CHRO and the second to VP of Talent Acquisition. A chief human resources officer runs the people function. A vice president of talent acquisition runs hiring. A benchmark, here, is a comparison of those jobs across companies. A candidate intake document is the form that starts a search. Second, build an agent around the customer’s own expertise. The organization puts its own method, standards, and decision rules into a Studio workflow. The customer supplies the method. Studio supplies the people intelligence, the execution, and the checks. Third, use Studio inside tools the team already has. The page names the Findem platform, Claude, ChatGPT, Microsoft Copilot, and Gemini. It also names Findem MCP, which brings Findem’s intelligence into the team’s own agents, applications, and workflows. MCP is the Model Context Protocol, a shared plug so one AI tool can read another system’s data. The Findem page prints the letters. It does not spell them on this post. Claude is Anthropic’s assistant. ChatGPT is OpenAI’s. Copilot, here, is Microsoft’s. Gemini is Google’s. The page says “compatible AI environments such as” that list. It does not say the list is closed. Do not add a product it left off. File the three paths as the company’s. This desk did not connect one.

The data number, and whose number it is. The section “The foundation behind work teams can trust” says trust starts before an agent receives a prompt. A prompt is what someone asked. Findem’s AI Labeling Engine turns scattered people data into structured context. Findem MCP then gives Studio agents access to the 3D People Graph’s 1.6 trillion expert-labeled data points across people, companies, and time. 1.6 trillion is 1,600 billion. It is the company’s count of those labeled points. A data point, here, is one tagged fact. The page does not define the 3 in 3D. The same sentence places those points across people, companies, and time. The about box later says the graph connects trillions of data points. Trillions is the about box’s word. 1.6 trillion is the body’s figure. This desk did not reconcile the two, and it did not count the points. The same section says the graph puts the right people, companies, and evidence in front of the model from the start. Agents then apply a defined method created by a named practitioner or supplied by the organization. The page says “named practitioner.” It does not print that person’s name on this post. Do not add one. The work that comes back is supposed to show the evidence, the criteria, and the reason for the conclusion. Agent outputs are recommendations for a team to review and refine. People make the final decisions. A recommendation is not a hire, and it is not a promotion. File that limit as the page’s.

Four quotes, as color, not as a measured result. Hari Kolam, co-founder and chief executive, said the next phase of enterprise AI is about completing meaningful work, not simply helping people find information. He said that takes more than a capable model. It takes trusted intelligence, deep expertise, and clear accountability for the result. He said people can now try Studio agents for free and see what that next phase looks like. Josh Bersin, named on the page as a global industry analyst and chief executive of The Josh Bersin Company, said HR has moved from asking whether AI can do the work to whether a team can trust it. He said people data must be accurate, current, and complete, and that Findem structured and labeled that data and had practitioners check it. He said that combination turns Findem’s people data into information an enterprise can trust. Casey Firey, named head of global talent operations at Camunda, said the Findem MCP agent they use inside Camunda is doing extraordinary work, that people are impressed when they show it, and that they wanted people intelligence inside the processes they already run. He said they have only scratched the surface. Camunda is the employer the page names. This desk did not watch that demo. Kyle Lagunas, named founder and principal of Kyle & Co., said talent-tech vendors have left data readiness and governance to the client, and that Findem took the opposite position with Studio. He said every agent carries a point of view on how the work should be done, drawn from practitioners who have done it well, and that a talent leader can read that point of view, agree with it, or change it. The page prints a stray comma in that sentence, “agree with, it.” This filing does not repeat the broken comma. He said that is a different start than a blank prompt, and that the last mile of AI is where products differ. File the names, the titles, and those sentences as the post’s. A quote is not a customer count, and it is not a score.

Where to try it, and what the about box adds. Findem Studio and Findem MCP are now available, on the page’s words. The free try path printed in the opening is https://studio.findem.ai/solutions. A later line says visit studio.findem.ai/solutions. The page does not print a price for a paid product. Do not add one. It does not name a customer who turned Studio on today, apart from Firey’s sentence about a Findem MCP agent already inside Camunda. That sentence is about the MCP agent. It is not a count of Studio launch customers. The about box calls Findem the AI infrastructure for people-centric work. It says the People Intelligence platform turns scattered people data into context that teams and their AI systems can reason over and act on. It says expert labeling turns that data into consistent, evidence-backed signals about experience, capabilities, and relationships, so a team can see why a conclusion was reached. Explainability is that last idea: a person can follow the reason. It says Findem’s agents and enterprise applications run on this foundation with human review built in, and that partners can build on the same base. It says Fortune and Fast Company named Findem one of the most innovative companies. Those names are the company’s. This desk did not open either list. It says Findem is trusted by industry leaders including FedEx, Intuit, RingCentral, and Emirates. Including means the page does not claim the list is complete, and it does not say those four bought Studio on 22 Sep. Do not turn the list into launch customers.

Plain English for the rest of the card: Findem Studio = the product launched on this page, agents that return finished people work. people intelligence = structured facts about people, companies, and time. agent = software that produces a piece of work, not only a chat reply. succession plan = who could fill a key job. market map = who works where in a field. hiring brief = the write-up of a role before a search. skills gap = abilities a team has versus abilities it needs. org chart = who reports to whom. expert-labeled = a person who knows the field tagged the fact. method = the written way the work is supposed to be done. CHRO = chief human resources officer. VP of Talent Acquisition = the executive who runs hiring. MCP = Model Context Protocol, the shared plug. The page prints the letters and does not spell them. 3D People Graph = Findem’s name for the data the agents read. 1.6 trillion = the body’s count of expert-labeled data points. Trillions, plural, is the about box’s word for the same graph. They are not two counts this desk merged. recommendation = the agent’s output. People make the final call. This filing is the 22 Sep launch. It is not a hire this desk watched.

PRIMARY here: Findem’s 22 Sep 2026 news post, “Findem Launches Studio: People Intelligence, Built for AI, That Returns Finished Work Teams Can Trust,” at findem.ai/news/findem-launches-studio-people-intelligence-platform — Tier A PRIMARY, the company’s own record. The dateline is Redwood City, Sept. 22, 2026. The page does not print an hour. The launch of Findem Studio, the finished-work list, the three requirements, the succession-plan example, the three ways to use it, the CHRO and talent-acquisition benchmarks, the Claude, ChatGPT, Copilot, Gemini, and Findem MCP names, the AI Labeling Engine, the 1.6 trillion data points, the named-practitioner line with no name printed, the recommendation limit, the Kolam, Bersin, Firey, and Lagunas quotes, the free path at studio.findem.ai/solutions, the available-now line for Studio and Findem MCP, the about-box trillions wording, the Fortune and Fast Company line, and the FedEx, Intuit, RingCentral, and Emirates “including” list are the post’s. NOT claimed: an hour stamp, a price, a named practitioner, that this desk ran an agent or counted 1.6 trillion points, that Firey’s Camunda sentence is a Studio launch-customer list, that FedEx, Intuit, RingCentral, or Emirates bought Studio on 22 Sep, that this desk checked the Fortune or Fast Company lists, a stock tip, or investment advice. Distinct from the already-filed perform-ai-commerce, rabbit-os3, and lumos-mcp-governance.