AI Is Rewriting Enterprise IT Integration

For decades, the backbone of global enterprise IT has relied on a manual, labor-intensive process known as system integration. Companies like SAP, Accenture, and Deloitte have built empires connecting disparate software systems, a task that often consumes years and hundreds of millions of dollars, with a staggering 70% failure rate. This foundational work, essential for everything from filling prescriptions to managing global supply chains, has remained largely untouched by automation – until now. The technology to fundamentally change this paradigm has finally arrived, as detailed in a recent a16z Blog post.

The complexity of legacy systems, often featuring millions of lines of custom code written by departed employees and inconsistent data schemas, has historically demanded human judgment and contextual reasoning. Basic concepts like 'customer' or 'order' can mean entirely different things across systems, making automated integration nearly impossible.

The Human Bottleneck

This inherent difficulty fostered a massive industry of consultants, embedding thousands within Fortune 500 companies. Software licenses alone account for billions annually, but the cost of making that software functional within an enterprise is several times higher, contributing to a global IT services spend nearing $1.8 trillion.

The delivery model has barely evolved since the early days of SAP. Large-scale migrations, like those involving SAP ECC, can stretch for years and cost upwards of $500 million. With SAP ending support for ECC in 2027, companies face immense pressure to upgrade, often opting to defer the problem due to the prohibitive cost and risk of failed transformation programs.

AI Enters the Fray

Kabir Nagrecha, founder of Tessera, argues that modern AI, particularly frontier models, possesses the capabilities to tackle these challenges. These models can now read obscure enterprise code like ABAP, map heterogeneous data schemas, and infer business logic from system behavior—tasks that previously necessitated costly human expertise.

This marks a significant shift, moving beyond simple data shuffling to genuine code comprehension and problem-solving. The long tail of judgment-heavy, structured work, once the domain of high-priced consultants, is becoming tractable for AI-driven agent systems.

System Integration as Software

Tessera is positioning itself as an AI enterprise transformation platform, where software performs the heavy lifting of integration. Humans are reserved for strategic decisions, architectural planning, and stakeholder alignment. Their initial focus is on ERP transformations, ingesting customer environments and automating the classification, migration, and validation processes for platforms like SAP S/4HANA migration.

A small Tessera team can now achieve what previously required dozens of consultants over two years. This approach fundamentally differs from 'autocomplete for consultants' tools, aiming instead to replace the human billable hour model with outcome-based software delivery.

Incumbent giants like Accenture and Deloitte face an innovator's dilemma; their business models are built on billable hours, making a shift to software-driven, outcome-based pricing a significant threat to their revenue streams.

The vision extends beyond ERPs to encompass the entire constellation of enterprise systems—CRMs, HRIS, supply chains—all requiring continuous integration and modernization. By transforming system integration as software, the cost of enterprise IT could fall by an order of magnitude, dramatically accelerating business adaptability.

Companies adopting this AI-driven model will gain a significant structural advantage. This represents the largest untapped market in enterprise software, poised to eliminate friction and boost productivity across economies.