1. A real-time fault repair recommendation system comprising: an input/output (I/O) interface configured to receive one or more multimodal user inputs 5 and a preconfigured prompt; a memory configured to store instructions; and a processor operatively coupled to the memory and the I/O interface and configured to: recognize one or more patterns indicative of at least one problem, from the received one or more multimodal user inputs; and 10 generate a unified context based on the one or more recognized patterns, using domain related information; the processor in combination with a Generative Artificial intelligence (Gen AI) model, using the preconfigured prompt, is further configured to: identify a core problem from the generated unified context; 15 decompose the identified core problem into one or more simplified subproblems and assign the one or more simplified sub-problems to respective one or more layers, wherein each layer contains at least one simplified sub-problem; dynamically allocate one or more tools to solve the one or more simplified sub-problems assigned at each layer; and 20 output a fault repair recommendation plan for the identified core problem by aggregating one or more solutions proposed by the one or more allocated tools.

2. The system as claimed in claim 1, wherein the identified core problem is decomposed into one or more simplified sub problems recursively, where the one or more simplified 25 subproblems present in one layer is decomposed to one or more sub problems in next layer until at least one of : all the one or more simplified sub problems are solved by the GenAI, a decomposition threshold is reached, and one or more selected tools are allocated to the one or more simplified sub problems is satisfied, wherein once one or more tools are allocated, the one or more simplified subproblems requires no further decomposition.

3. The system as claimed in claim 1, wherein for dynamic tool selection, the processor is configured to: collate one or more tools and metadata associated with the one or more tools for addressing the one or more simplified sub problems; and 5 assist the GenAI model based on the one or more collated tools and their associated metadata and at least one of historical data, machine logs, error codes, manual, components of a machine to be examined for selection of one or more tools.

4. The system as claimed in claim 1, wherein the one or more tools to solve the one or more 10 simplified sub-problems is implemented using at least one of: traditional Artificial Intelligence, programmable API, Generative Artificial Intelligence, or a combination thereof.

5. The system as claimed in claim 1, wherein the processor coupled to the memory is further 15 configured to: synthesize a seamless repair recommendation solution by aggregating the one or more solutions proposed by the one or more tools, post tool execution, wherein the seamless repair recommendation solution is synthesized by parsing the one or more solutions using Traditional AI tools or GenAI or a combination thereof. 20

6. The system as claimed in claim 1, wherein the processor coupled to the memory is further configured to: provide the repair recommendation plan to a user for the identified core problem; receive a feedback from the user in response to execution of the repair 25 recommendation plan; and store the feedback in the memory to improve accuracy of the repair recommendation plan.

7. The system as claimed in claim 1, wherein the domain related information comprises associations between abbreviations and domain specific terms present in the recognized patterns to generate the unified context. 5 8. The system as claimed in claim 1, wherein the one or more user inputs is received from the group consisting of: typed input; audio input; video input; 10 audiovisual input; and input provided via a graphical user interface.

9. The system as claimed in claim 1, wherein the I/O interface is further configured to receive a timestamp information along with the received one or more multimodal user inputs. 15

10. The system as claimed in claim 1, wherein the one or more patterns is recognized using at least one of Natural Language Processing (NLP) technique, Generative Artificial Intelligence (Gen AI) technique or a combination thereof

Description:AS FILED PDF DOCUMENT , Claims:AS FILED PDF DOCUMENT

I/We Claim:
1. A real-time fault repair recommendation system comprising:
an input/output (I/O) interface configured to receive one or more multimodal user inputs
5 and a preconfigured prompt;
a memory configured to store instructions; and
a processor operatively coupled to the memory and the I/O interface and configured to:
recognize one or more patterns indicative of at least one problem, from the received
one or more multimodal user inputs; and
10 generate a unified context based on the one or more recognized patterns, using
domain related information;
the processor in combination with a Generative Artificial intelligence (Gen AI) model,
using the preconfigured prompt, is further configured to:
identify a core problem from the generated unified context;
15 decompose the identified core problem into one or more simplified
subproblems and assign the one or more simplified sub-problems to respective one
or more layers, wherein each layer contains at least one simplified sub-problem;
dynamically allocate one or more tools to solve the one or more simplified
sub-problems assigned at each layer; and
20 output a fault repair recommendation plan for the identified core problem
by aggregating one or more solutions proposed by the one or more allocated tools.
2. The system as claimed in claim 1, wherein the identified core problem is decomposed into
one or more simplified sub problems recursively, where the one or more simplified
25 subproblems present in one layer is decomposed to one or more sub problems in next layer
until at least one of : all the one or more simplified sub problems are solved by the GenAI,
a decomposition threshold is reached, and one or more selected tools are allocated to the
one or more simplified sub problems is satisfied, wherein once one or more tools are
allocated, the one or more simplified subproblems requires no further decomposition.

3. The system as claimed in claim 1, wherein for dynamic tool selection, the processor is
configured to:
collate one or more tools and metadata associated with the one or more tools for
addressing the one or more simplified sub problems; and
5 assist the GenAI model based on the one or more collated tools and their associated
metadata and at least one of historical data, machine logs, error codes, manual, components of
a machine to be examined for selection of one or more tools.
4. The system as claimed in claim 1, wherein the one or more tools to solve the one or more
10 simplified sub-problems is implemented using at least one of: traditional Artificial
Intelligence, programmable API, Generative Artificial Intelligence, or a combination
thereof.
5. The system as claimed in claim 1, wherein the processor coupled to the memory is further
15 configured to:
synthesize a seamless repair recommendation solution by aggregating the one or
more solutions proposed by the one or more tools, post tool execution, wherein the
seamless repair recommendation solution is synthesized by parsing the one or more
solutions using Traditional AI tools or GenAI or a combination thereof.
20
6. The system as claimed in claim 1, wherein the processor coupled to the memory is further
configured to:
provide the repair recommendation plan to a user for the identified core problem;
receive a feedback from the user in response to execution of the repair
25 recommendation plan; and
store the feedback in the memory to improve accuracy of the repair
recommendation plan.

7. The system as claimed in claim 1, wherein the domain related information comprises
associations between abbreviations and domain specific terms present in the recognized
patterns to generate the unified context.
5 8. The system as claimed in claim 1, wherein the one or more user inputs is received from the
group consisting of:
typed input;
audio input;
video input;
10 audiovisual input; and
input provided via a graphical user interface.
9. The system as claimed in claim 1, wherein the I/O interface is further configured to receive
a timestamp information along with the received one or more multimodal user inputs.
15
10. The system as claimed in claim 1, wherein the one or more patterns is recognized using at
least one of Natural Language Processing (NLP) technique, Generative Artificial
Intelligence (Gen AI) technique or a combination thereof