The modern corporation is often bogged down by its own administrative weight — a friction that persists despite decades of digitization. Patrik Bergareche Sáiz de los Terreros, who previously led the food delivery giant Just Eat, is betting that artificial intelligence can finally dismantle these bureaucratic hurdles. His new venture, Punto, has secured €2 million in seed funding from the venture capital firm Samaipata to bring that thesis to market.

Punto aims to streamline the repetitive, often manual processes that clog corporate workflows. By deploying AI to handle administrative tasks, the startup seeks to move beyond simple automation toward a more autonomous operational layer. The focus is not merely on speed, but on the systematic reduction of the "hidden work" that consumes employee productivity and stalls organizational momentum.

The back office as the next frontier

For most of the software era, corporate back offices have been digitized rather than truly automated. Enterprise resource planning systems, expense platforms, and HR tools moved paper processes onto screens, but the underlying logic — approvals, reconciliations, compliance checks, data entry across disconnected systems — still required human hands at nearly every step. The result is a persistent layer of low-value labor that scales roughly in proportion to headcount, creating drag that grows alongside the organization it serves.

The arrival of generative AI and large language models has shifted the calculus. Tasks that once resisted automation because they involved unstructured data, contextual judgment, or natural-language communication are now within reach of software agents. A wave of startups across Europe and the United States has begun targeting this space, each carving out a niche: invoice processing, contract review, procurement workflows, employee onboarding paperwork. Punto appears to position itself not as a point solution for one of these verticals but as a broader platform for back-office autonomy — an ambitious framing that will eventually need to be tested against the specificity that enterprise buyers tend to demand.

Samaipata, the Madrid- and London-based fund backing the round, has built a portfolio concentrated in European marketplace and platform businesses. Its willingness to back an AI-native operations startup signals a thesis that the next generation of enterprise value creation may come less from new consumer-facing platforms and more from the invisible infrastructure that keeps existing businesses running.

From logistics to administrative logic

Bergareche's trajectory from Just Eat to Punto is worth examining on its own terms. Food delivery platforms are, beneath their consumer interfaces, orchestration engines: they coordinate merchants, riders, payments, regulatory compliance, and customer service across thousands of simultaneous transactions. The operational discipline required to run such a system at scale shares structural similarities with the challenge of managing a large company's administrative backbone, where dozens of processes must interlock reliably despite fragmented tooling.

That said, the enterprise back office presents its own distinct obstacles. Corporate buyers move slowly, procurement cycles are long, and integration with legacy systems is often the real product challenge rather than the AI model itself. Many promising automation startups have discovered that the technology works in demos but stalls during deployment, caught between the messiness of real-world data and the rigidity of existing IT architectures. Whether Punto can navigate this gap will depend less on the sophistication of its models and more on the quality of its implementation playbook.

The broader market context adds both tailwind and noise. Enterprise spending on AI tooling is accelerating, but so is vendor fatigue. Chief information officers report being inundated with pitches from startups claiming to automate one process or another, making differentiation increasingly difficult. The startups that gain traction tend to be those that can demonstrate measurable cost reduction or time savings within weeks, not quarters.

Punto enters this environment with a credible founder pedigree and early capital, but at a stage where the product's actual capabilities and go-to-market strategy remain largely undefined in public view. The tension worth watching is a familiar one in enterprise AI: the distance between a compelling thesis about eliminating bureaucratic friction and the granular, often unglamorous work of making that thesis survive contact with a real company's workflows.

With reporting from El Confidencial.

Source · El Confidencial — Tech