# US 2026/0094100 A1 · Systems and methods for automating cloud architecture optimization using a generative artificial intelligence orchestrator

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0094100 A1
- Document ID: US20260094100A1
- Status: Published application
- Kind: A1
- Application number: 18/898,800
- Filed: 2024-09-27
- Published: 2026-04-02
- Earliest claimed: 2024-09-27
- Assignees: AT&T Intellectual Property I, L.P.
- Applicants: AT&T Intellectual Property I, L.P.
- Inventors: Venkateswarlu Maddi; Anthony Snead; Narendra Avineni; Christopher Hambrick; Rama Raghavan
- Web page: https://patentgrep.com/patent/US20260094100A1

## CPC

- G06Q 10/06375: {Prediction of business process outcome or impact based on a proposed change}
- G06F 9/5072: {Grid computing}
- G06F 9/5077: {Logical partitioning of resources; Management or configuration of virtualized resources (specific details on emulation or internal functioning of virtual machines G06F9/455)}
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Aspects of the subject disclosure may include, for example, running a virtual agent to initiate retrieval of information about cloud subscriptions and cloud resource properties; connecting the virtual agent with a generative artificial intelligence (GenAI) server via a first application programming interface (API) and a second API; with the first API, parsing the information about cloud subscriptions and cloud resource properties and generating a prompt for cost saving recommendations; with a second API, filtering the prompt and sending the prompt that complies with security policies to the GenAI server; and with the GenAI server, generating cloud architecture optimization (AO) recommendations in response to the prompt. Other embodiments are disclosed.

## Claims

### Claim 1

A device, comprising:
a processing system including a processor; and
a memory that stores executable instructions that, when executed by the processing system, facilitate performance of operations, the operations comprising:
running a virtual agent to initiate retrieval of information about cloud subscriptions and cloud resource properties;
connecting the virtual agent with a generative artificial intelligence (GenAI) server via a first application programming interface (API) and a second API;
with the first API, parsing the information about cloud subscriptions and cloud resource properties and generating a prompt for cost saving recommendations;
with a second API, filtering the prompt and sending the prompt that complies with security policies to the GenAI server; and
with the GenAI server, generating cloud architecture optimization (AO) recommendations in response to the prompt.

### Claim 2 (depends on 1)

The device of claim 1, wherein the operations further comprise:
receiving cloud AO guidelines by AO architects via a knowledge base portal; and
storing the cloud AO guidelines in a vector database; and
wherein the generating the cloud AO recommendations further comprises, with the GenAI server, generating the AO recommendations by performing a cognitive search in the vector database.

### Claim 3 (depends on 1)

The device of claim 1, wherein the operations further comprise training the GenAI server with a set of training data using input parameters retrieved from the cloud subscriptions and the cloud resource properties.

### Claim 4 (depends on 1)

The device of claim 1, wherein the operations further comprise:
presenting, with the virtual agent, guided queries to users to retrieve information relating to a target application migrated into a cloud.

### Claim 5 (depends on 4)

The device of claim 4, wherein the running the virtual agent further comprises running a chatbot application to retrieve the information relating to the target application including at least identification information of the target application.

### Claim 6 (depends on 4)

The device of claim 4, wherein the operations further comprise:
with the first API, receiving, from the virtual agent, the cloud resource properties of the target application and parsing the cloud resource properties of the target application.

### Claim 7 (depends on 1)

The device of claim 1, wherein the filtering the prompt further comprises:
determining whether the prompt contains sensitive personal information, payment information, proprietary information, profane words, or a combination thereof.

### Claim 8 (depends on 1)

The device of claim 1, wherein the operations further comprise:
with a second API, rejecting the prompt that fails to comply with security policies and sending an error message to the first API,
wherein the first API is a server API and the second API is a GenAI API.

### Claim 9

A non-transitory machine-readable medium, comprising executable instructions that, when executed by a processing system including a processor, facilitate performance of operations, the operations comprising:
receiving, via a virtual agent, a cloud architecture optimization (AO) request with respect to a target application migrated into a public cloud subscribed by a customer enterprise;
calling, by the virtual agent, a server API to parse an input object containing cloud resource properties owned by a subscription of the target application and generating a prompt using templates;
invoking a generative artificial intelligence API (GenAI API) which performs security policy validations of the prompt;
upon passing of the security policy validations of the prompt, sending the prompt to a GenAI server; and
generating, by the GenAI server, cloud architecture optimization (AO) recommendations.

### Claim 10 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the operations further comprise invoking, by the virtual agent, a client API to retrieve the cloud resource properties owned by the subscription of the target application.

### Claim 11 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the operations further comprise:
receiving, via a batch process, the AO request with respect to a group of applications scanned by the batch process and migrated into the public cloud subscribed by the customer enterprise; and
invoking, via the batch process, a client API to retrieve cloud resource properties relevant to the group of applications.

### Claim 12 (depends on 11)

The non-transitory machine-readable medium of claim 11, wherein the operations further comprise:
calling, by the batch process, the server API to parse input objects containing the cloud resource properties relevant to the group of applications and to generate another prompt using the templates; and
invoking the GenAI API which performs the security policy validations of another prompt.

### Claim 13 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the generating the cloud AO recommendations comprises generating the cloud AO recommendations by comparing the prompt with AO guidelines maintained in a vector database using a cognitive search,
wherein the cloud AO recommendations are directed to cost savings by performing resource consolidation, resources optimization, logging optimization, Platform as a Service (PaaS) optimization and PaaS change, converting Infrastructure as a Service (IaaS) to PaaS/Software as a Service (SaaS), or a combination thereof.

### Claim 14 (depends on 9)

The non-transitory machine-readable medium of claim 9, wherein the operations further comprises:
generating, with the server API, a first prompt that instructs the GenAI server to output a structured query language (SQL) query;
invoking, with the server API, a particular program code library;
generating, with the server API, a second prompt that instructs the GenAI server to output particular program code using the SQL query; and
calling, with the server API, the particular program code library to execute the particular program code, thereby generating the AO recommendations.

### Claim 15 (depends on 14)

The non-transitory machine-readable medium of claim 14, wherein the first prompt comprises a first instruction to generate the SQL query using rules IDs, and the second prompt comprises a second instruction to generate the particular program code using the generated SQL query.

### Claim 16

A method, comprising:
receiving, by a processing system of a generative artificial intelligence (GenAI) orchestrator including a processor, identification of a target cloud application migrated into a public cloud subscribed by a customer enterprise;
establishing connections, by the processing system, among a server API, a generative artificial intelligence API (GenAI API), and a virtual agent or a batch process;
automatically generating, by the processing system, using the server API, a prompt based on a plurality of rules, wherein the prompt is configured to instruct generation of cloud architecture optimization (AO) recommendations directed to cost savings with respect to the target cloud application;
sending, by the processing system, the prompt to a generative artificial intelligence (GenAI) server via the GenAI API; and
returning, by the processing system, the generated cloud AO recommendations to the virtual agent or the batch process.

### Claim 17 (depends on 16)

The method of claim 16, further comprising:
receiving, by the processing system, subscription properties of the target cloud application retrieved by a client API via the virtual agent or the batch process; and
parsing, by the processing system, using the server API, the subscription properties of the target cloud application to generate the prompt.

### Claim 18 (depends on 16)

The method of claim 16, further comprising:
calling, by the processing system, a particular program code library; and
connecting, by the processing system, using the server API, the particular program code library and the GenAI API to instruct the GenAI server to generate a particular program code.

### Claim 19 (depends on 18)

The method of claim 18, further comprising:
calling, by the processing system, using the server API, a particular program code library to execute the generated particular program code; and
generating, by the processing system, the AO recommendations based on the executed particular program code.

### Claim 20 (depends on 16)

The method of claim 16, comprising:
facilitating, by the processing system, connections with user interfaces supporting multiple channels configured to provide a set of parameters relating to the target cloud application in different formats, wherein the multiple channels comprise the virtual agent and the batch process.
