Tesla is actively developing a feature that will allow drivers to give Full Self-Driving (Supervised) highly specific, natural-language parking instructions. The confirmation comes directly from Ashok Elluswamy, Tesla’s VP of AI Software, who recently revealed that the team is tackling one of the most common friction points of the FSD experience.
In a recent exchange on X, Tesla owner Chris outlined a frequent frustration with using FSD in residential areas. He pointed out that the software would be significantly more useful if drivers could simply tell the car which driveway to pull into, much like they would direct a human driver.
Chris highlighted that traditional GPS pins are notoriously inaccurate for specific houses, and that Tesla currently lacks a way to input visual context. He pitched the idea of telling the car, “It’s the white house on the left, just past that SUV,” and having FSD remember that exact drop-off preference for future trips.
Ashok Elluswamy confirms development
The concept quickly caught the attention of Tesla’s AI leadership. Elluswamy replied directly to the post with a brief but definitive confirmation: “Working on it.”.

This brief statement officially validates that Tesla is building a conversational bridge between its in-car AI, Grok, and the vehicle’s driving behavior. Instead of manually taking over the steering wheel when FSD hesitates at a destination, drivers will soon be able to use their voice to guide the final parking maneuver.
The Grok and FSD roadmap
Tesla has been steadily expanding the role of Grok inside its vehicles since late 2025. What began as a conversational assistant has evolved into a tool for complex route planning and navigation. Integrating Grok into FSD’s planning layer is Tesla’s next software architecture leap.
This aligns closely with recent comments from CEO Elon Musk, who has teased that spoken FSD commands are slated for a fall 2026 release. Musk has previously highlighted upcoming features like destination parking memory and a reverse-summon “Banish” feature, which perfectly complement the user request Elluswamy just acknowledged.
The final 30 seconds of a trip are often the most unpredictable for daily FSD users. The system frequently defaults to the first available spot or stops awkwardly in the street rather than pulling neatly into a driveway. By enabling users to verbally specify their precise drop-off location, Tesla is bringing FSD closer to resembling a genuine robotic chauffeur.
While Elluswamy didn’t reveal a specific release date, his confirmation indicates that Tesla is actively considering community feedback regarding edge-case maneuvers.

