Sidekick runs factory work orders and training through plain text messages

Why it matters

Sidekick is testing whether AI workplace software can reach frontline employees by removing the app entirely, while taking on the accuracy and access-control risks of factory operations.

Justin So (@Just1nSo) detailed Sidekick's frontline operations product in an August 7th post on X, pitching plain text messages as the interface for reporting equipment failures, finding procedures and routing factory work.

https://x.com/Just1nSo/status/2085787245110583594

Sidekick gives workers a phone number rather than a workplace app. They can send questions, photos and voice memos, including messages in languages other than English. Sidekick searches an employer's documents, replies over SMS and escalates unanswered questions to a manager.

The San Francisco business is part of Y Combinator's Summer 2026 batch. YC lists So as Sidekick's founder and CEO, with a master's degree in computer science from the University of Pennsylvania and prior experience in finance at chip designer Arm. YC's directory currently describes Sidekick as a one-person team.

So has also described Sidekick as a product he built with a co-founder. YC partner Charlie Warren identified Aadip Khanal as So's co-founder in a post about the business. Warren said he backed the pair through YC.

Built from the founder's factory access

Sidekick grew out of So's exposure to his father's manufacturing operations. In an earlier announcement, So said his father owns and operates manufacturing facilities and was repeatedly interrupted when workers could not find answers in existing documents or systems.

That connection gave Sidekick an initial testing ground and a specific distribution path. So said Sidekick began on his father's factory floor before expanding to other sites. He currently claims the product is operating in more than 20 plants and serving over 600 frontline workers. Those adoption figures are self-reported and Sidekick has not named the plants.

The worker-facing interaction remains deliberately narrow. Employees do not create an account or remember a password. Managers use a web dashboard to load documents, review activity and manage the information that Sidekick can retrieve.

Sidekick's broader product reaches beyond question answering. When a worker reports a problem, Sidekick says it can classify the issue, create a work order, route it to the appropriate employee and preserve the resolution. Sidekick also maintains equipment records containing service history, manuals and maintenance information.

The same SMS channel supports employee training. Sidekick's website shows managers assigning short modules, videos and quizzes for topics such as chemical handling, forklift certification and lockout procedures. The goal is to bring training into a channel workers already use instead of requiring another piece of software on the factory floor.

Every resolved problem becomes data

Sidekick's central product bet is that a stream of routine questions can become an operational record. A supervisor's response to a machinery problem can be stored, cited and returned when another worker encounters the same issue. Over time, Sidekick aims to build a searchable history of repairs, procedures and informal fixes that would otherwise remain in individual employees' memories.

That model also creates the most consequential product risk. Factory questions can involve maintenance, certification and safety procedures, where an incorrect answer carries greater consequences than a mistaken response from a general workplace chatbot.

Sidekick's terms say its AI-generated responses may contain errors and place responsibility on users to verify information before relying on it. Sidekick says its answers are grounded in employer-provided material, and its website presents source citations as part of generated reports. The quality of those answers will still depend on the accuracy of the underlying documents and the controls surrounding high-risk instructions.

SMS removes much of the rollout burden that has slowed workplace software adoption among employees who spend their shifts away from desks. It also gives Sidekick a larger job to handle behind the conversation: authenticating workers, limiting access to sensitive documents and deciding when a request should reach a human manager.

So is betting that factories will accept those tradeoffs to reach workers through an interface already present on virtually every phone. Sidekick's early plant deployments give him a base for testing that thesis. The harder task is proving the system can turn informal factory knowledge into dependable instructions without introducing a new source of operational error.