Tapestry and PJM outline AI pilot to speed interconnection reviews

Tapestry, an Alphabet Moonshot lab initiative, and PJM on day two of the FERC Software Conference described a collaboration to apply artificial intelligence to speed power‑grid interconnection reviews. Kat Wong, head of technical operations at Tapestry, said the project will create a unified data model and a "Knowledge Graph" for the electric grid to surface asset status and flag inconsistencies in application materials. "We are making the electric grid visible and intelligent," she said.

PJM interconnection planning manager AJ Lambert said PJM has shifted from first‑come, first‑served to a cluster‑based, readiness‑focused tariff and is implementing a transition plan that will put many applications through synchronized decision points. Lambert said PJM and Tapestry will pilot automated document verification and automated model validation to help the operator manage hundreds to thousands of applications under the new process. "Our ultimate vision is to get to an inbox‑zero — clearing the backlog of old projects — while making interconnection predictable and timely," Lambert said.

Why it matters: Interconnection delays are widely cited as a bottleneck for new generation and storage. PJM has regulatory approval for a cluster‑based approach but faces operational challenges: synchronizing reporting for many projects, validating readiness documentation and ensuring model inputs are technically correct. Tapestry’s approach is intended to reduce manual review time while surfacing only the items that require human engineering judgment.

Details and limitations: Presenters described a phased workplan that begins with automated cross‑checks for consistency across filing elements (form fields, one‑line diagrams, uploaded models) and moves toward automated checks of modeled parameters (e.g., dynamic model flags) and simulation benchmarking against PJM’s tools. They emphasized the need for curated, high‑quality training data, and multiple participants asked about benchmarking and verification; Tapestry said benchmarking is ongoing and that simulation tools on the platform are being used to compare results with PJM expectations.

Next steps: The teams said the first operational tool to support PJM transition points is planned later this year, with later phases expanding to broader planning and collaborative workflows. PJM stressed continued human oversight for tariff compliance and study interpretations; Tapestry emphasized that automation is intended to reduce workload, not to replace engineering judgment.