This test result made me really very happy earlier today 😎

We've been testing AI photo identification accuracy - especially on equipment with no brand markings (text / logos), which is very common with technical equipment out in the field.

We chose Texecom's Capture MIR to do the tests, and they results were as follows:

• ChatGPT was worst, giving a very generic answer (i.e. saying it was a 'PIR sensor', but no model);

• Google Gemini was equal worst, as it gave the completely wrong model (a Bosch PIR);

• Super Engineer WON - giving a 100% accurate answer (Texecom Capture MIR).

The interesting thing about Super Engineer's answer is that this specific model has two versions - a PIR and an MIR, both of which look identical in form BUT which have different coloured LEDs that light up.

In this case Super Engineer spotted not only the correct model type (a Texecom Capture sensor) but ALSO that it was the MIR version, due to the purple LED.

That was VERY SMART of Super Engineer. Congrats to our AI team for achieving this.

See full test results below . . .

Super Engineer answer = 100% correct

Chat GPT answer = 100% general

Google Gemini answer = 100% wrong