Circuit Mind introduces automated schematic verification

London-based Circuit Mind has introduced AVE, an automated verification application that compares electronic circuit schematics against component datasheets, application notes and internal engineering requirements. The system uses specialised AI agents to examine electrical connections and operating constraints, with a further quality-assurance process reviewing identified issues before they are presented to engineers.

The Agentic Verification Engine extends Circuit Mind’s existing electronics design platform, which already includes the ACE circuit-generation application. While ACE produces electronic schematics and supporting information from engineering requirements, AVE examines completed designs for potential errors and deviations from their specified operating conditions.

Electronic schematic verification requires engineers to compare component connections with information distributed across datasheets and application documentation. A single device may have several supply pins, optional interface configurations and restrictions governing voltage, current, timing or power sequencing.

Those requirements become more difficult to track when a circuit incorporates hundreds of components. Information may be divided among electrical-characteristics tables, pin descriptions, absolute-maximum ratings and application notes, with no common documentation structure across manufacturers.

Conventional electrical rule checking can detect problems such as incompatible pin types or certain unconnected signals, according to the rules and component definitions supplied to the design environment. It cannot be assumed to verify every operating condition described in the component manufacturer’s documentation.

AVE extends the review process by examining information from those documents alongside the schematic. Some agents analyse pin tables, supply voltages and other component requirements, while others trace connections and buses across the circuit to identify relationships that need further examination.

A voltage connection, for example, may appear electrically valid within a schematic editor but still exceed a component’s recommended operating range. Similarly, connecting two digital devices may require attention to their supported signalling levels, interface configuration or additional external circuitry.

Detecting these conditions requires the review system to associate documentation with the correct component and its actual connections. It must also distinguish absolute-maximum ratings from recommended operating values, since a circuit operating below a damage threshold is not necessarily functioning within its specified conditions.

Circuit Mind has developed structured digital models of electronic components to support this interpretation. The platform converts information from manufacturer documentation into machine-readable records, reducing dependence on the layout and wording of individual datasheets when its applications examine component requirements.

The system can also incorporate an organisation’s engineering context, including approved manufacturers, preferred components, internal checklists and design rules. These requirements may be more restrictive than the component’s published electrical limits, particularly where a company has established its own derating, qualification or supply-chain policies.

Before findings are returned to the engineer, AVE performs a further quality-assurance pass intended to challenge and filter the results produced by its specialist agents. The company says this process reduces unsupported alerts, although it has not published an independently verified detection rate or false-positive benchmark for the new application.

Reported issues are associated with the relevant component, circuit connection and supporting documentation. That allows engineers to inspect the evidence before deciding whether the schematic requires modification, rather than relying solely on an automated conclusion without access to its underlying technical basis.

The application is intended to work with existing electronic computer-aided design environments. Schematics created within other ECAD tools can be reviewed through the platform, while ACE-generated designs can use the same verification process. Circuit Mind identifies integration with established engineering systems as part of its approach rather than requiring teams to replace their existing design tools.

ACE performs a different engineering function. It takes a high-level hardware architecture and specified requirements, selects components, performs circuit calculations and produces schematics, bills of materials and supporting analysis. Its design options can be evaluated against constraints including power consumption, cost, physical dimensions and component availability.

Verification remains necessary even when a circuit has been generated automatically. The interpretation of requirements, accuracy of component records and assumptions underlying a design must still be checked against the equipment’s intended operating conditions before physical construction.

Nor can schematic verification establish every characteristic of a finished printed circuit board. Layout-related signal integrity, electromagnetic compatibility, thermal behaviour, manufacturing tolerances and physical reliability may require separate analysis, simulation and testing. AVE’s announcement concerns schematic review rather than comprehensive validation of fabricated hardware.

The application arrives alongside a $25m financing disclosure, bringing Circuit Mind’s reported total funding to $33.4m. The financing includes an $18.5m Series A led by Norwest, together with a previously undisclosed seed round led by Third Prime.

Circuit Mind intends to use the investment to extend its platform into further development activities, including analysis, reporting, testing and firmware. It also plans to expand its engineering and commercial operations and establish a headquarters in the United States.

The company identifies Schneider Electric and Emerson among organisations working with its wider technology. Schneider Electric has reported early evaluations suggesting substantial reductions in electronic architecture exploration effort, although those results relate to the broader design platform and do not constitute a performance benchmark for AVE.

AVE is available through Circuit Mind’s platform for customer evaluation. Its contribution to engineering workflows will depend on the quality of the underlying component information, the completeness of internal design rules and the accuracy of the issues identified for review before hardware manufacture.