A smarter border usually brings to mind trucks and cargo at a checkpoint, with scanners reading containers, biometric kiosks identifying travelers, and AI flagging anomalies. But for entrepreneurs the relevant border is really three: one for goods, one for services, and one for talent. Some ventures need the physical movement of products, so trade facilitation remains central; others depend on cloud infrastructure, AI platforms, payment systems, and data flows that never touch a checkpoint yet are governed by regulation on both sides; still others need access to talent, through visas, daily commuter flows in the border region, or binational remote teams. And all three are shifting into digital space, where decisions increasingly take place in databases, and are made by models, before any product or person reaches a physical checkpoint. That physical-digital reversal, if properly designed, can make the smarter border welcoming for smaller firms that today bear a disproportionate share of the cross-border operational cost, and will also determine whether entrepreneurship can emerge across borders at all: ventures born binational, enabled by the same digital infrastructure the smarter border runs on.

What follows is an argument that how those upstream decisions are governed—the data that agencies collect, and the models that act on it—will determine whether cross-border entrepreneurship remains a privilege of well-funded or large firms that can absorb the compliance costs, or an opportunity for talent to create value across the Mexico-US border. Cross-border ventures depend on three flows—goods, services, and talent—and each reaches the border at a different layer: goods at the checkpoint, services at the layer of data flows, and talent at the identity layer. A smarter border has to work at all three at once, and artificial intelligence is what makes that feasible, because the bottleneck in each layer is the same one: how two jurisdictions agree on what counts as a trustworthy shipment, a compliant data flow, or a credentialed worker.

The digital shift this depends on is already underway: shipments, travelers, and firms are routinely cleared or flagged in data before they reach the physical line, so the digital representation of a movement now precedes and shapes the movement itself. The mistake is to read this as a digitization problem—putting yesterday’s paper forms into today’s digital systems. The real work is to redesign how information moves and how decisions about it are governed, so that agencies, entrepreneurs, brokers, and cross-border talent become participants in the same ecosystem rather than parties on opposite sides of a counter.

I discuss below four governance principles. They are not about the technology, which will keep changing, but about how authority, information, and trust are arranged across the new layer, which is what determines who can build a cross-border business and who cannot.

1.        Level the playing field between small and large firms

Trusted-trader status today (CTPAT in the United States, OEA in Mexico) carries real benefits, but the operating history, documented security and personnel standards, and approval timelines required to obtain it put it out of reach for almost every startup. The old framing treats facilitation and control as a tradeoff that trusted trader programs were designed to manage; in the digital layer the tradeoff loosens, because risk models can recalibrate continuously from information rather than rationing scrutiny by category, so trust can be assessed as a moving signal rather than a one-time certification.

Trust can also come from outside the government, through private verification networks that already attest to identity, credit, employment, and compliance history in other regulated domains, and that the border layer can read from instead of rebuilding from scratch. If trust becomes machine-readable and portable across agencies and countries, a small logistics startup in Tijuana, Mexico, can earn a verifiable risk profile through its transaction history rather than through an out-of-reach certification. The same logic applies to talent, where portable and verifiable credentials would let a software engineer with five years of remote work for US firms prove that record once and have it recognized on both sides, instead of starting from zero with every new visa or contract. Identity should remain distinct from risk, so that young ventures and unproven workers can earn trust transaction by transaction rather than being kept out by requirements they cannot yet meet.

2.        Convert border uncertainty into predictable cost

For an entrepreneur, the worst feature of a border is not friction but variance. A shipment that might clear in two hours or two days makes inventory planning, customer expectations, and working capital impossible to manage, and a single missed delivery can end a customer relationship that took months to build. When clearance decisions happen in data before goods move, the border becomes predictable, and a small firm in Monterrey can credibly quote a two-day window to a Texas buyer instead of pricing in a week of safety stock. Compliance then stops being paperwork to file and becomes infrastructure to build on; manifests, certificates of origin, and security attestations move upstream into the software that runs the business, and the entrepreneur experiences the border as an API rather than a checkpoint. The same shift matters in the other two layers: in services, predictable rules on cross-border data flows let a fintech in Guadalajara design a product around US users without redesigning compliance every time the rules shift; in talent, predictable mobility and remote-work rules let a binational team plan a year of work rather than a quarter.

3.        Make automated decisions explainable and contestable

A large firm can absorb an unexplained algorithmic flag because it already has lawyers on retainer and brokers with direct lines to the agency; a small exporter facing the same flag, with no path to challenge it, may simply exit the market. Audit trails, human review, and appeal mechanisms are important market-entry conditions, and the design choice is about who gets to operate at the border at all. Human review matters for a second reason as well: a reviewer who sees a confident, well-formatted AI output is far less likely to question it than one who sees the same conclusion arrived at through messy human reasoning, and the result is a safeguard that exists on paper but is rarely exercised in practice. Explainability without a human empowered and incentivized to override is therefore close to no safeguard at all. The same logic argues for collecting less data in the first place. Every additional field, attestation, or registration is a cost that falls hardest on firms without an IT department, and the principle of minimum-necessary collection is as much about market access as it is about data privacy and cybersecurity exposure.

4.        Treat shared infrastructure that entrepreneurs can build on

Single windows like Mexico’s VUCEM (Foreign Trade Single Window) and the United States’ Automated Commercial Environment (ACE), digital identity systems, and open data standards are the foundation on which the rest of the cross-border economy gets built. When governments keep this layer interoperable, openly documented, and free from lock-in to any single vendor, they make space for a secondary ecosystem to form on top: a Tijuana-based compliance startup that turns customs filings into a self-serve product for small exporters, a fintech that settles binational invoices in hours rather than days, a credential platform that lets a project manager carry verified work history across the line. The same governments that today try to build closed end-to-end systems would do better to build narrow, well-documented interfaces and let entrepreneurs build the surfaces that face users. That is how public infrastructure usually wins: GPS, the Internet, and ACH each began as government-supplied foundations that became platforms only because they were open enough for others to build on.

Melvin Kranzberg observed that “technology is neither good nor bad; nor is it neutral,” and a smarter border will be exactly what the institutions and organizations building it decide to make it. Deploying AI at the border rearranges who decides, on what evidence, with what review, and on whose behalf—inside agencies and inside firms. Treating the change as a technical upgrade misses what actually shifts: not the tools, but the allocation of authority. The institutions and entrepreneurs that thrive in the coming decade will treat the border as a continuously operating information system, governed with accountability and human oversight, where compliance becomes infrastructure, decisions can be contested, and the cost of crossing stops scaling with firm size. That is what makes cross-border entrepreneurship viable for the firms the current system effectively keeps out, and it is the shift I think is worth building toward.

If you build ventures across borders, regulate those who do, or think about how those rules should change, I would like to hear where this argument holds, where it should be extended, and where it misses cases I have not seen.