# US 2026/0018159 A1 · Telecommunications switch-type infrastructure for communications with call router and audio record server for computational source-to-target language conversion via application of artificial intelligence agents

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

- Number: US 2026/0018159 A1
- Document ID: US20260018159A1
- Status: Published application
- Kind: A1
- Application number: 19/261,984
- Filed: 2025-07-07
- Published: 2026-01-15
- Earliest claimed: 2024-07-12
- Assignees: Cheryl Ee-Lin Cunningham
- Inventors: CHERYL EE-LIN CUNNINGHAM; JAMES DAVID HARLOW
- Web page: https://patentgrep.com/patent/US20260018159A1

## CPC

- G10L 13/00: Speech synthesis; Text to speech systems
- G10L 15/16: using artificial neural networks
- H04L 51/02: using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A class-4 telecommunications switch hosts artificial-intelligence agents that convert live speech to text, translate the text between languages, and synthesize natural speech in real time. The switch proxies calls among public trunks, PBX/media gateways, and cloud ACD/CRM services, embedding diacritic-rich transcripts and user-specific language-model personalization. Deployable at the customer edge, in the PSTN core, or as SaaS, the system supports one-to-one, one-to-many, many-to-one, and many-to-many call patterns. FPGA, ASIC, or SoC accelerators minimize latency and bandwidth, cutting capital cost while improving global voice interoperability and cybersecurity.

## Claims

### Claim 1

An apparatus comprising:
a telecommunication switch-type infrastructure including one or more processing units configured to execute artificial intelligence (AI) processes, neural network processes, or other computer-readable instructions directed to computational linguistics, wherein said infrastructure is connected to an automatic call dispatcher and an audio record server, located between them or within a call center, and further comprising: (a) a Class-4 software-defined networking softswitch configured in a Computational Interpretation of Communication using Artificial Intelligence (CICAI) deployment; (b) one or more interpretation, transcription, translation, or transliteration operations, or any combination thereof, performed on digital signals representative of one or more telecommunication sessions; (c) one or more AI agents performing said operations for respective users in real-time or near real-time; (d) processing units including one or more FPGAs, ASICs, or SoCs dedicated at least in part to computational linguistics including semiological or hermeneutical interpretation of human language; (e) said infrastructure operating in one or more peer-to-peer, one-to-many, many-to-one, or many-to-many configurations; and (f) personalization of said operations based on a user's vernacular.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the transcription includes insertion of diacritics to mimic intonation, pauses, or vowel pronunciation.

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein the telecommunication switch-type infrastructure is located within 10 meters of a call center demarcation point.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein the infrastructure communicates using a TCP/IP signaling protocol or equivalent packet-based protocol.

### Claim 5 (depends on 1)

The apparatus of claim 1, wherein an incoming call is received via a T1/T3 PSTN/POTS line at a PBX/media gateway and directed to a call center computing device via a CTI/ACD server, with CRM data provided to the computing device via a cloud CRM system.

### Claim 6 (depends on 1)

The apparatus of claim 1, wherein said one or more AI agents include a first AI agent representing an originating caller and a second AI agent representing a destination caller, each configured to perform their respective AI operations.

### Claim 7

A method comprising:
performing, via a telecommunication switch-type infrastructure having one or more processing units, AI processes, neural network processes, or other computer-readable instructions directed to computational linguistics; wherein the method further includes: (a) receiving incoming calls via PSTN/POTS lines, (b) directing such calls via a PBX/media gateway and CTI/ACD server to user devices, (c) interpreting, transcribing, translating, or transliterating signals using said infrastructure in real-time or near real-time using AI agents, (d) customizing operations to the user's vernacular, and (e) using CRM systems to personalize and guide routing decisions.

### Claim 8 (depends on 7)

The method of claim 7, further comprising inserting diacritics into transcriptions to reproduce intonation or pauses.

### Claim 9 (depends on 7)

The method of claim 7, wherein said AI agents include separate agents for the originating and receiving parties.

### Claim 10 (depends on 7)

The method of claim 7, further comprising operating the infrastructure in peer-to-peer, one-to-many, many-to-one, or many-to-many mode.

### Claim 11

An article comprising: a non-transitory storage medium having stored thereon instructions executable by a special-purpose computing platform to: (a) perform, via a telecommunication switch-type infrastructure, AI and neural network processes directed to computational linguistics; (b) communicate with call dispatch and record servers; (c) interpret, transcribe, translate, or transliterate digital signals between communication endpoints in real-time or near real-time using AI agents.

### Claim 12 (depends on 11)

The article of claim 11, wherein the instructions include logic for operating with Class-4 software-defined CICAI softswitch functionality.

### Claim 13 (depends on 11)

The article of claim 11, wherein transcription includes diacritics representing intonation, pauses, or phonetic variation.

### Claim 14 (depends on 11)

The article of claim 11, wherein CRM integration is provided to inform call routing and personalization.

### Claim 15 (depends on 11)

The article of claim 11, wherein the instructions include logic to deploy AI agents specific to caller endpoints.

### Claim 16 (depends on 11)

The article of claim 11, wherein said instructions perform customization based on vernacular using updated user data.

### Claim 17 (depends on 11)

The article of claim 11, wherein one or more processing units comprise FPGAs, ASICs, or SoCs.

### Claim 18 (depends on 11)

The article of claim 11, further comprising logic for controlling junction points between circuit-switched and packet-switched networks.
