The least glamorous job in enterprise technology may also be one of its most expensive: being the person between two systems that refuse to agree. A new employee appears in one database but not another. A laptop is marked returned in one tool and active in the next. A license stays attached to someone who left months ago. A ticket opens, a queue grows, and a capable human becomes a courier for context.
Mayan Mathen has been watching versions of this scene for more than a quarter century. He has viewed it from network rooms in Johannesburg, global technology meetings at Dimension Data, an applied research lab at NTT in Silicon Valley, a connected-home startup, and the corporate-development side of an acquisition. His current company, XOPS, is his attempt to remove the courier work without removing the people.
The proposition is simple enough to fit on a Post-it and difficult enough to occupy a career: people should declare the outcome, software should coordinate the systems, and humans should supervise what happens. Tickets are evidence that work needs doing. Mathen wants the machinery to finish it.
The student who preferred the data center
Mathen's path into technology began at the University of the Witwatersrand, where he later recalled being more interested in the IT department than going to class. He built things in the university data center and campus network. Someone at a networking specialist noticed. That company eventually became part of Bytes Technology Group, where he spent five years learning the practical grammar of infrastructure.
In 2004 he moved to Dimension Data for what he called a broader view of the technology landscape. The move delivered exactly that, plus a punishing travel schedule. He worked in converged communications, business development and technology leadership, helping open new territories and develop new business units. Colleagues called him the “chief travelling officer.” At seven feet tall, he was also hard to miss: Mathen joked that strangers on his trips sometimes asked to be photographed with him. Somewhere, by his own reckoning, he appears in holiday albums owned by people he never met again.
The distinction is useful. Mathen's public persona has room for jokes, Formula One, aircraft and the delight of a flaming sambuca at a company function. His work, though, moved toward larger systems and less forgiving consequences. By the time he became CTO for Dimension Data's Middle East and Africa operations, he was spending much of his life with clients, vendors, analysts and executives. The technical problem was rarely just technical. New markets demanded different models, different assumptions and often different products.
A laboratory for moving quickly
In 2013 Mathen came to Silicon Valley to help establish NTT Innovation Institute, an applied R&D and customer experience center. As SVP and CTO, he worked across cloud computing, security, mobility, software-defined networking and network-function virtualization. The assignment joined two habits that still shape his thinking: collaborate widely, then make something concrete before the meeting notes go cold.
At an NTT forum in 2014, he described how a coffee conversation with a friend at Plantronics about wearable technology led toward prototype work. The institute's rhythm was to build quickly, test whether an idea worked, ask whether operating companies could sell it, and then improve the product alongside customers. His line on innovation was relational rather than mystical: “Collaboration and innovation go together very closely.”
He could also be cheerfully brutal. When a technology writer showed him a framework for bringing mobile computing into the enterprise, Mathen's verdict was: “CIOs will love it. It will never work.” The problem, he argued, was IT's inherited control mentality. Mobile devices, networks and software changed too rapidly to fit a world of centrally chosen, locked-down standards. Trying to extend the old control model would make IT less relevant to the business it served.
That observation reads now like an early draft of the XOPS thesis. Complex systems do not become manageable because a department insists on more control. They become manageable when the operating model accepts complexity, maintains reliable context and gives people a better place to stand.
The founder loopConnected homes, an acquisition, then back to enterprise IT
Mathen left NTT in 2015 to found Catalina Labs. The company worked on the connected home, using data from networks and devices to help consumers understand their digital lives and to help service providers support them. The object changed from enterprise infrastructure to the home, but the fascination was familiar: many devices, incomplete visibility, and a person trying to make sense of the whole.
Catalina Labs was acquired by Asurion. Mathen moved into the larger company as vice president of corporate development, where his work included launching services for major enterprise clients. He had now seen the founder's sketch, the operator's deployment and the acquirer's strategic logic. By April 2022, he was ready to start again.
Human role: supervise, judge exceptions and improve the policy. System role: keep working until reality matches the declared outcome.
With co-founder Andrew Robinson, Mathen built XOPS around what the company calls an active system of intelligence. The platform keeps a living knowledge graph across the enterprise and uses goal-oriented software robots to act across systems. Instead of writing a brittle path for every possible situation, a company declares the result it needs: provision this employee, recover that device, reclaim this license, enforce this policy. The platform coordinates the steps and checks the result.
The emphasis on autonomy invites the usual question about reliability. XOPS's answer is unusually stern for the AI era: execution must be deterministic, auditable, replayable and reversible. Reasoning can help decide. The platform layer cannot roll dice. In a consumer chatbot, a surprising answer may be charming. In access control or asset recovery, charm is an operational defect.
Out of stealthProduction before the press release
XOPS remained quiet for three years. When it emerged from stealth in August 2025, the company announced $40 million in funding co-led by Activant Capital and FPV Ventures. It also said the platform was already running in Fortune 500 environments. The order mattered. Enterprise software buyers are professionally suspicious, and rightly so. A demo can be beautiful while a global asset register is still a swamp.
One disclosed customer result gives the abstraction some edges. XOPS reported that a large enterprise reduced staffing needs on a major annual initiative by 95 percent, shortened a process from days to under an hour, and reached 100 percent asset-tracking accuracy. The customer was not named, and the figures are company-reported. Still, they clarify the kind of work at stake: repetitive coordination at a scale where small mistakes multiply.
The current company figures widen the frame. XOPS says it coordinates more than 1.4 million employees, orchestrates $4.8 billion in IT spending and is deployed across 39 countries. The software is technical. The aspiration is almost domestic: fewer queues, fewer loose ends, fewer Mondays spent asking which database is telling the truth.
Serious work, lightly worn
There is a pleasant tension in Mathen's story. He likes speed, experimentation and people who refuse to think inside the box. He also now argues for operational machinery that behaves predictably every time. Both instincts make sense when placed in sequence. Explore freely while discovering the product. Execute precisely when the product controls business-critical work.
The same balance appears in his temperament. The young technologist who preferred a campus network to class became a global ambassador, a CTO, a founder, an acquired founder, and a founder again. He remained the very tall Formula One fan who enjoyed making colleagues laugh. Personality was not sacrificed to seriousness. It made seriousness easier to carry.
XOPS is still a private company making ambitious claims in a crowded category. Its long-term verdict will arrive through customers, renewals and the unromantic evidence of systems behaving correctly. Mathen, at least, has chosen a problem proportionate to his experience. Every stage of his career put him near a boundary: university and industry, network and application, R&D and commercialization, startup and acquirer, one enterprise system and another.
Boundaries are where the courier work accumulates. They are also where a patient operator can see the next company hiding in plain sight. Mathen's bet is that enterprise IT no longer needs another place to describe the work. It needs a dependable way to bring the work to an end.