To currently recommend sensors requires quite a lot of domain knowledge and human phone calls. To speed this up, one can use AI in two smart ways. 

Create a knowledge base that is filled from the spec and data sheets from products from manufacturers. This can in the end by something specific like a Spotify webshop for example (does not need be customer facing itself). 

Create an AI tool that actually uses this knowledge base to answer questions from customers and asks questions with respect to the context in which the lighting or beyond lighting system is being used. For example, a customer might ask: "What PIR sensor works for this building with a height of 20 meters?" On top of that, the AI might actually ask for the height of a particular building if the customer fails to provide this information.  

Note that the AI tool might be used by support staff at the company, rather than it being directly used by the end-customer. In that way it would increase the quality of the responses of the support staff without exposing the end-customer to potentially hallucinating AI answers. 

[1] The Role of Generative AI in E-commerce (Rasińska, 2024) 

Compensation: This internship is paid.