The documented record on this topic is thin for a European audience by name — the sourced case studies run through Toronto, Redwood City, Tokyo and Silicon Valley, not Frankfurt or Paris. But the two companies the wider industry names as Europe's platform-scale entrants into agentic AI governance, Mistral and Siemens, sit inside a race being driven by a single, widely circulated statistic that Via News's own verification layer cannot stand behind.
The number driving the rush — and a reason to doubt it
According to survey data reported by MIT Technology Review, within two years 100% of respondents plan to be using agentic AI, with 69% expecting to use it widely1. The same reporting puts average enterprise AI agent access to company data at 45%, falling to 30% or less at organizations categorized as "data laggards"1. Two-thirds of those laggards — 66% — say legacy data systems limit how far they can scale agents, and 68% say the same systems prevent agents from deciding at speed, against just 8% of "data leader" firms reporting either constraint1.
Those numbers are the ones circulating in enterprise AI conversation right now, and they are the ones underpinning the general claim that adoption intent has outrun data readiness. Here is the seam worth showing rather than hiding: when Via News's fidelity checks ran against this MIT Technology Review piece, 0% of the 11 checked claims held up against primary sourcing. That does not mean the figures are wrong — surveys of this kind are frequently paraphrased or re-aggregated in ways that break a strict fidelity check — but it means readers should treat the headline percentages as directional signal, not settled fact, until a primary source is independently confirmed.
Where the European names actually appear
Mistral and Siemens are named among the large platform players — alongside Microsoft and NVIDIA — said to be embedding agentic AI into governance and engineering workflows through enterprise partnerships. The dossier behind this piece, however, carries no documented case, quote, or figure for either company's specific initiative. The best-evidenced version of that pattern instead comes from North America: Manulife, the Toronto-headquartered insurer, has renewed and expanded a five-year partnership with Microsoft, adopting the Microsoft Frontier Suite and deploying Microsoft Agent 365, alongside expanding Microsoft 365 Copilot to more than 30,000 employees2. "Our partnership with Microsoft is a critical enabler of Manulife's continued evolution into a truly AI-driven organization," said Shamus Weiland of Manulife, describing the Frontier Suite adoption as giving the insurer a foundation to advance AI across its global operations with the confidence, security, stability and intelligence needed to embed AI responsibly2. That announcement, like the Box disclosure below, is a company press release; Via News's reliability tracking on this news wire shows only 33% of 2,547 previously checked claims held up, a caveat worth carrying alongside the specifics rather than treating them as independently confirmed.
A parallel governance push appears at Box, whose new controls — agent guardrails, third-party agent activity oversight, prompt injection detection, and classification-based access policies — are aimed at letting enterprises deploy AI agents against their content with more confidence3. Nomura Research Institute's Tatsutoshi Murata welcomed the move: "As we rapidly advance our utilization of AI agents, we expect Box — which has consistently led the development of security management capabilities for secure collaboration — to provide the administrative features needed to safely leverage this new era of AI," he said, adding that Nomura values Box's multi-vendor flexibility across AI models and is confident in its protective layer as agent use expands3. The same 33%-of-2,547 reliability caveat applies here.
The regulated-industry logic that speaks directly to European back offices
The part of this dossier that translates most directly to a European business audience is not the platform-partnership layer but the vertical-specialist one — because the logic vertical AI companies use to pick their first markets is the same logic that governs regulated European sectors. Maisa AI's CEO, David Villalon, frames the company's market as "the market of process automation of core business and production tasks at regulated industries," adding: "So, at the end, it's all the core tasks that are today being manual or handled by humans that are part of the core product or the core services of the company."4 Per the dossier's characterization of that market, the work in question spans back office, operations and finance — precisely the manual, auditable, reproducible, hallucination-resistant processes that European financial-services and healthcare compliance regimes already demand be traceable, whoever performs them.
The sizing case for one such vertical comes from healthcare administration in the US, offered by Penguin AI's Glenn Herzberg: "US Healthcare Administration runs about a trillion dollars a year, about a quarter of the total health spend, and the published estimates put around $570 billion of that in work that has no effect on health outcomes."5 Penguin AI's framing — administrative labor spend, not IT software budget — is itself the strategic point: it targets a cost line large enough to fund a company without needing broad access to a client's entire data estate, exactly bypassing the 45%-data-access ceiling described above (with the reliability caveat on that figure noted).
A second example, further from healthcare but the same shape, is Casap, a dispute-automation company for fintechs. Its founders, Shanti and Sayisi, "had both experienced the pain points of disputes firsthand at their respective large fintech companies and saw like the amount of internal effort and organizational work that it took to solve those challenges even with a really strong engineering team," according to Primary partner Emily Man6. Man said Primary was drawn to the founders' backgrounds and the problem they were pursuing — "We were immediately really excited about them because of their backgrounds, and they talked to us about this opportunity that they were thinking about tackling" — and described feedback from people who had worked with them as "resoundingly clear that these were two exceptional builders who were really passionate about starting their own thing and solving problems that they had seen before."6
What to watch
For European readers, three threads are worth tracking rather than assuming settled. First, whether Mistral's or Siemens's specific agentic AI governance moves generate documentable, checkable announcements — the dossier behind this piece has none yet, only their inclusion in the general trend. Second, whether the 45%-average and 30%-laggard data-access figures survive independent primary-source confirmation, given the 0%-of-11 fidelity result on the reporting that carries them. Third, whether the vertical-specialist pattern — Maisa in regulated process automation, Penguin AI in healthcare administration, Casap in fintech disputes — reaches European regulated industries with the same auditability pitch, since that pitch, unlike the platform-partnership headlines, is not contingent on unresolved enterprise data-access constraints.
