Best Fit
Teams running multi-constraint board programs, such as industrial control, embedded and connected devices, power and mixed-signal hardware, that need faster iteration with traceable decisions.
AI-Assisted PCB Design
From Engineering Brief to Review-Ready PCB Layout.
@I Design turns a written engineering brief into a complete PCB design package (schematic, netlist, placed-and-routed layout, BOM, DRC report and Gerber files) for teams that need shorter design cycles without giving up review control.
AI does the drafting: it interprets the brief, proposes components and explores layout trade-offs. A deterministic rule engine decides what is acceptable, enforcing electrical safety, manufacturing limits and your own constraints before anything reaches your reviewers.
- Plain-language brief in; schematic, layout, BOM and Gerbers out.
- Low-confidence interpretations are flagged for human review, not guessed.
- Private beta with selective, small-batch intake.
How It Works
AI Proposes. Rules Decide. Engineers Approve.
@I Design is an AI-assisted PCB design system that combines a domain-trained language model, electronic design automation modules and a deterministic rule engine in one pipeline.
Every stage reads from and writes to one shared design representation, so each component choice and layout decision stays traceable to the requirement that caused it.
- Brief to intent: requirements, limits and implied constraints extracted, each with a confidence score.
- Intent to components: parts ranked on electrical fit, cost, availability and reliability.
- Components to schematic: netlist and schematic assembled from proven circuit patterns.
- Schematic to layout: placement and routing optimised for trace length, vias, EMI, thermal load and signal integrity.
- Layout to release: DRC report, BOM and Gerbers handed to your engineers for sign-off.
Fit Boundary
Who This Is For and What It Is Not For
The strongest fit is an engineering team where design-cycle time and review quality both carry real cost, with schedule, compliance or reliability consequences attached to each board.
Not A Fit
Open-ended experimentation, one-off drafting requests, or programs that cannot state requirements, acceptance criteria and a review cadence. The system accelerates engineering work; it does not replace engineering sign-off.
Review Control
Built to Be Reviewed, Not Trusted Blindly
Every design passes explicit checkpoints where the system either shows that a rule is met or hands the decision to a person.
Ambiguous requirements trigger a question, not an assumption. When an AI proposal breaks a rule, the system relaxes only within safe margins, tries an alternative approach, falls back to a proven circuit pattern, or stops and flags the issue for an engineer. Safety-related rules, such as isolation, clearance and thermal limits, hold a veto that optimisation cannot override.
- Confidence scoring on every extracted requirement.
- Staged fallback that ends in mandatory human review.
- A decision log recording each rule intervention and the alternatives considered.
Status
Private Beta. Patent Pending.
The system behind @I Design is the subject of New Zealand patent application 820555.
Access is limited to a small number of engineering programs at a time, so each engagement gets architect-level attention. Benchmark figures are published on the validation page together with the method used and its limits.
- Patent pending: NZ application 820555.
- Architect-led collaboration from intake to handoff.
- Benchmark figures published with method and limits.