Granted Patent B2

US 12,524,709 · App. 18/759,267 · Granted Jan 13, 2026

Inventors: Julie J. Brown (Yardley, PA); Young M. Lee (Old Westbury, NY); Rajiv Ramanasankaran (San Jose, CA); Sastry KM Malladi (Fremont, CA); Michael Tenbrock (Dachsen, CH); Levent Tinaz (Tampa Bay, FL); Samuel A Girard (Kenosha, WI); David S. Elario (Hartland, WI); Juliet A Pagliaro Herman (Waukesha, WI); Miguel Galvez (Westford, MA); Trent M. Swanson (Wellington, FL); John F. Kuchler (Muskego, WI); Deepak Budhiraja (Ashburn, VA); Daniela M. Natali (Kensington, MD); Josip Lazarevski (Zurich, CH); Scott Deering (Milwaukee, WI); Gary W. Gavin (Franklin, WI); Kristen Sheppard-Guzelaydin (West Chester, PA); James Young (Cork, IE); Prashanthi Sudhakar (San Francisco, CA); Kaleb Luedtke (West Bend, WI); Karl F. Reichenberger (Mequon, WI); Wenwen Zhao (Santa Clara, CA); Adam R. Grabowski (Brookfield, WI); Lauren C. Dern (Fox Point, WI); Nicole A. Madison (Milwaukee, WI); Dana S. Petersen (Milwaukee, WI); Nevin L. Forry (York, PA); Pedriant Pena (Groveland, MA); Ghassan R. Hamoudeh (San Marcos, CA); Ryan G Danielson (Castle Rock, CO)

Assignee: TYCO FIRE & SECURITY GMBH

G06N20/00G05B23/0243G05B23/0275G06N3/0475

Abstract

Claims (42)

1 . A system, comprising:

one or more processors to:

receive a prompt indicating a request for servicing an item of building equipment;

apply the request as input to a generative artificial intelligence (AI) model comprising at least one generative pre-trained transformer, to cause the at least one generative pre-trained transformer to generate information including a prediction of a root cause of a problem associated with the request for servicing the item of building equipment, the generative AI model configured based on training data that comprises a plurality of examples of requests for servicing building equipment and a plurality of examples of root causes of problems associated with the plurality of examples of requests;

generate steps for responding to the problem; and

update the information in response to an evaluation of user engagement relating to performance of the steps for responding to the problem.

2 . The system of claim 1 , wherein the training data comprises natural language data representing the plurality of examples of requests and the plurality of examples of root causes.

3 . The system of claim 1 , wherein the training data comprises data from a fault detection and diagnosis (FDD) model.

4 . The system of claim 1 , wherein the one or more processors are to receive the prompt via a conversational interface, the prompt comprising at least one of text, audio, speech, image, or video data.

5 . The system of claim 1 , wherein the one or more processors are to receive feedback regarding the prediction and update the prediction, using the generative AI model, based on the feedback.

6 . The system of claim 1 , wherein the one or more processors are to generate, using the generative AI model, a service report comprising the prediction and conforming to a predetermined format for the service report.

7 . The system of claim 1 , wherein the one or more processors are to receive feedback regarding the prediction and update the generative AI model based on the feedback.

8 . The system of claim 1 , wherein one or more processors are to generate, using the generative AI model, based on the prediction, a recommendation to adjust control of the item of building equipment.

9 . The system of claim 1 , wherein the one or more processors are to:

identify one or more first parameters of the generative AI model to maintain at predetermined values; and

perform a training process to configure one or more second model parameters using feedback.

10 . The system of claim 1 , wherein:

the steps for responding to the problem indicate a type of service activity; and

the evaluation of user engagement relates to performance of the type of the service activity.

11 . The system of claim 1 , wherein the one or more processors are to (i) apply the request as input to the generative artificial intelligence (AI) model comprising the at least one generative pre-trained transformer, to cause the at least one generative pre-trained transformer to generate the information including the prediction of the root cause of the problem and (ii) update the information, in an application session for the user.

12 . A system, comprising:

one or more processors to:

receive a prompt indicating a problem of an item of building equipment;

apply the problem as input to at least one neural network configured as a generative pre-trained transformer, to cause the at least one generative pre-trained transformer to generate information including a prediction of a root cause of the problem, the at least one neural network configured based on training data that comprises a plurality of examples of requests for servicing building equipment and a plurality of examples of root causes of problems associated with the plurality of examples of requests; and

update the information based on whether a user discontinued providing inputs to the system.

13 . The system of claim 12 , wherein the plurality of examples of root causes of problems comprises data from a fault detection and diagnosis (FDD) model.

14 . The system of claim 12 , wherein the one or more processors are to receive the prompt via a conversational interface, the prompt comprising at least one of text, audio, speech, image, or video data.

15 . D) The system of claim 12 , wherein the at least one neural network is coupled with a data source regarding the item of building equipment to retrieve sensor data regarding the item of building equipment and generate the prediction further based on the sensor data.

16 . The system of claim 12 , wherein the one or more processors are to receive feedback regarding the prediction and update the prediction, using the at least one neural network, based on the feedback.

17 . The system of claim 12 , wherein the one or more processors are to generate, using the at least one neural network, a service report comprising the prediction and conforming to a predetermined format for the service report.

18 . The system of claim 12 , wherein the one or more processors are to receive feedback regarding the prediction and update the at least one neural network based on the feedback.

19 . A method, comprising:

receiving, by one or more processors, a prompt indicating a request for servicing an item of building equipment;

applying, by the one or more processors, the request as input to a generative artificial intelligence (AI) model comprising at least one generative pre-trained transformer, to cause the at least one generative pre-trained transformer to generate information including a prediction of a root cause of a problem associated with the request for servicing the item of building equipment, the generative AI model configured based on training data that comprises a plurality of examples of requests for servicing building equipment and a plurality of examples of root causes of problems associated with the plurality of examples of requests;

generating steps to respond to the problem; and

updating the information in response to an evaluation of user engagement with an application relating to:

performance of the steps to respond to the problem; or

discontinuation of providing inputs to the application.

20 . The method of claim 19 , further comprising:

receiving, by the one or more processors via conversational interface, feedback regarding the prediction;

updating the prediction, by the one or more processors, using the generative AI model; and

updating the generative AI model, by the one or more processors, based on the feedback.

Assignments (2)

Continuity (4)

References Cited (222)