Having 150,000 employees does not mean knowing who can do what

When companies say they lack talent, they usually look outside: job postings, search firms and competitors. A company with roughly 150,000 employees has another question to ask first. Is the capability truly absent, or is the organization unable to find people it already has?

On September 10, 2026, Mitsubishi Electric announced plans to build a global talent information platform by fiscal 2029 using talent data for roughly 150,000 group employees. AI agents are intended to analyze HR-specific unstructured data, convert it into forms that can be compared on common measures and support development, placement and acquisition decisions.

BAKUTAN, the platform provider, described a scope of more than 200 group companies and roughly 150,000 employees. This does not mean every record is already unified today. The public state is a platform build through fiscal 2029 with early use-case validation underway.

A smaller May experiment explains why the September platform matters

The September announcement is more informative when read alongside a May change. Mitsubishi Electric already had a mechanism for reconnecting people and jobs inside the group.

Its L.E.A.D program has managed a pool of roughly 300 next-generation management candidates. For fiscal 2026 the company introduced Proactive Frontier Posting, opening selected section-manager and department-manager positions to this pool. The postings use job descriptions and disclose information including job grade, compensation level, required capability and role difficulty.

The official May release also said Mitsubishi Electric planned a trial mechanism using AI to validate matching based on job descriptions and applicant HR information. In September, the company and BAKUTAN said they were testing the feasibility of an AI agent that supports applicant-to-job matching in this internal posting program.

January had already introduced mobility across national and organizational borders

The same direction appeared in January with Mitsubishi Electric’s Talent Mobility program. It targets mid-career and younger employees at overseas group companies and creates job matching across countries, regions and locations from fiscal 2026.

The company explicitly noted that information on strong mid-career and younger employees at overseas affiliates often remained utilized mainly within each local entity. It wanted to identify and share such talent earlier at group level and create borderless career opportunities.

Put January, May and September in sequence and the direction becomes clearer: owning talent is not enough. A large organization needs the ability to discover, compare and move capability across internal boundaries.

Talent scarcity inside a large company can partly be a search problem

Current title and department are easy to query. Experience launching a foreign business, working with a specific customer type, integrating organizations, or handling technology and operations together is much harder to discover across a complex group.

The larger and more international the group becomes, the more likely similar experience is recorded under different job titles, evaluation systems and document formats. A person may exist inside the company but remain functionally invisible to the team that needs that capability.

That is a very different use of AI from automatically rejecting applicants. The role here is to make fragmented, unstructured talent information more comparable and improve discovery between jobs and people.

Does better internal search mean less external hiring?

Better internal discovery could reduce some external hiring if the right people were already inside but hidden by information fragmentation. The current public materials, however, do not show an actual reduction in external hiring.

The opposite implication is also possible. Better visibility into internal capability can make the missing capability more precise. A broad request such as “we need AI talent” might become “we have model experience, but nobody who has integrated AI with industrial equipment in customer operations.”

That is Banseog’s inference, not a result claimed by Mitsubishi Electric. For executive search, however, the distinction matters: a more legible internal talent market may not eliminate external search so much as raise the precision required of it.

The useful HR AI test is not “can it analyze data?” but “does talent actually move?”

A dataset covering 150,000 employees is impressive, but volume alone is not strategic advantage. If the records remain incompatible or disconnected from real jobs and opportunities, a large talent pool can also be a large collection of hard-to-search documents.

What makes the Mitsubishi Electric case notable is the intended downstream use: development, placement and acquisition, with internal management-job matching among the early feasibility cases.

That suggests a more demanding way to evaluate HR AI. Do not stop at whether it can analyze employee data. Ask whether the analysis connects to real jobs, mobility, placement and development opportunities. At that point AI becomes infrastructure for reconnecting work and people inside the enterprise.

BANSEOG VIEW

Banseog View — Before asking “who should we hire?”, ask “is this capability really absent inside?”

Mitsubishi Electric’s January Talent Mobility program, May Proactive Frontier Posting and September global talent information platform are separate initiatives, but they point in a common direction: making internal experience and capability easier to discover and connect to real opportunities.

The evidence does not support saying that all 150,000 employees are already being automatically matched in a finished AI talent marketplace. What is public today is a platform plan through fiscal 2029 and feasibility testing in a limited internal job-posting use case.

The broader question is still useful. Before buying talent outside, how accurately can a company search what it already has and identify what is genuinely missing? As internal search improves, external search may increasingly become a more exact capability-acquisition problem.

SOURCES

Primary sources and references

Banseog does not describe the roughly 150,000-person dataset as already fully integrated. The global platform is planned through fiscal 2029, and the AI job-person matching currently confirmed is a feasibility exercise in an internal posting program for next-generation management candidates. The idea that better internal search could sharpen the boundary of external hiring is Banseog analysis, not a demonstrated Mitsubishi Electric outcome.