Minicor raises $3.8M to connect AI agents with desktop software

Minicor has secured $3.8 million in seed funding to connect AI agents with enterprise software that lacks application programming interfaces, or APIs, the company announced.

Brad Gillespie of IA Ventures headed the financing round, which also saw Y Combinator and its co-founder Paul Graham take part. Additional investors comprise Pioneer Fund, Robinhood Ventures, Garage Capital, Rebel Fund, Blast, SuccessVP, K5 Global, 20VC, Phosphor Capital, Metsu Capital, plus a number of private backers. Founded by Saheed Akinbile and Faizaan Chishtie, Minicor is developing a platform that enables AI agents to work with desktop programs as if they offered APIs.

The filing states that Minicor believes roughly 70 % of Fortune 500 firms continue to rely on desktop applications lacking APIs, which forces organizations to decide how to embed AI into current processes. One path involves robotic process automation (RPA), demanding engineers craft and upkeep task-specific scripts. A second alternative employs computer-use agents that manipulate software via its graphical user interface. While such agents are often simpler to roll out, they can operate more sluggishly than people and may falter when screens or procedures are altered.

Minicor’s strategy diverges by converting desktop program capabilities into callable endpoints for AI agents. According to the firm, its agents discover how each application operates, generate the required automations and keep them functional while the program is active. The goal is to link AI with the existing tools companies rely on instead of compelling them to swap out those solutions initially.

Practical Implications

Legacy desktop programs often underpin vital tasks, store important information, or sit at the heart of workflows that have evolved over decades. Re-engineering those procedures to revolve around modern software tends to be expensive and disruptive.

The success of Minicor’s method hinges on its ability to consistently manage a variety of applications and adjust as they evolve. Mistakes may prove especially expensive if agents deal with confidential data or execute actions that affect finance or operations. Companies will also seek proof that the solution cuts engineering effort while not creating additional upkeep issues.