XSparks Names Cosmo Mariano Chief Client Outcomes Officer to Turn AI Spending Into Margin for CEOs

XSparks, a global AI transformation firm, named Cosmo Mariano as Chief Client Outcomes Officer. A transformation leader with more than two decades scaling and transforming technology businesses and customer operations, Mariano joins the firm to close the gap between AI pilots and real business results.

Mariano leads client AI business strategy across the XSparks portfolio. His counterpart, CTO Angad Singh Wadhwa, leads AI technology strategy. Together their teams solve complex business problems with AI that works in production, then prove it on a number the CEO can take to the board.

The mandate targets the hours and headcount a business spends to operate software instead of running the business. What Cosmo calls “the software tax.” Workers log into ten systems and spend the day operating them: hunting for data, copying it from one app to the next, and updating records by hand, just to get the output their job needs. The cost is measurable. Knowledge workers switch between applications nearly 1,200 times a day and spend just under four hours a week reorienting after each switch (Harvard Business Review, 2022). That lost time compounds across an entire workforce. Asana’s research puts about 60% of the average knowledge worker’s day on work about work rather than the job itself.

Companies have spent heavily on AI. Enterprise spending on generative AI more than tripled in a year, from $11.5 billion to $37 billion (Menlo Ventures, 2025), yet 56% of CEOs report no financial benefit and only 12% captured both revenue and cost gains (PwC, 2026).

XSparks sees the deeper reason as structural. The first wave of enterprise AI was built to assist people, not to run the work: copilots that retrieve, summarize, and suggest while a person still operates the software by hand. The procedures underneath were never redesigned, so the work never left the people. Individuals got faster at their tasks. The cost of running the business stayed, and the margin never moved.

Also Read: AiThority Interview with Matej Bukovinski, Chief Technology Officer at Nutrient

“People blame the AI. The AI was never the problem,” said Mariano. “Companies added it on top of the way they already work, so the procedures never changed and the work never left their people. The tools sped individuals up, but the business ran the same way and the P&L never moved. And the pilots that did show promise died before production, because a pilot is not an operating model. Getting there takes redesigning the workflow, standing up the infrastructure to run AI on it, and keeping people in the loop to operate it. That is what we build. AI does the work. People direct it and own the outcome. You keep the margin.”

XSparks rebuilds a company so AI runs the work, rather than bolting AI onto legacy operations as another tool. The firm delivers through one methodology, Think. Build. Operate.: find where AI moves the P&L, deliver a first working system in four to six weeks, then operate and improve it after launch. Every engagement is guided by the AI Operating Model, or AIOM: the model XSparks deploys across consulting, technology composition, and managed operations. Its technology pillar is a seven-layer architecture that connects to the tools, data, and workflows a company already runs, so AI can move across the operation instead of stalling in a single pilot. XSparks reports the result to one quarterly number, the AI Return Multiple, measured across cost, revenue, time, capacity, quality, and risk.

Mariano’s work spans three fronts of the rebuild: business model design, product innovation, and the change management and education programs that carry a workforce through it. He helps CEOs redesign how the business makes money around what AI now makes possible, shapes the new AI-native products that come out of the rebuild, and builds the enablement that turns a plan on paper into a team that can run it.