DCT
1:21-cv-01698
Valyant Ai Inc v. Kea Cloud Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Valyant AI Inc (Colorado)
- Defendant: Kea Cloud Inc (Delaware)
- Plaintiff’s Counsel: Ashby & Geddes; Haynes & Boone LLP
- Case Identification: 1:21-cv-01698, D. Del., 02/09/2022
- Venue Allegations: Venue is alleged to be proper in the District of Delaware because Defendant is a Delaware corporation, has a registered agent in the state, and conducts business within the district.
- Core Dispute: Plaintiff alleges that Defendant’s conversational AI ordering system for restaurants infringes a patent related to an artificially intelligent order processing system that includes a human auditor for handling exceptions.
- Technical Context: The technology operates in the artificial intelligence sector, specifically applying natural language processing and machine learning to automate customer ordering in the quick service restaurant (QSR) industry.
- Key Procedural History: Plaintiff alleges it sent multiple cease-and-desist letters to Defendant between June and October 2021, providing Defendant with actual knowledge of the patent-in-suit prior to the filing of the complaint.
Case Timeline
| Date | Event |
|---|---|
| 2017-03-29 | ’706 Patent Priority Date |
| 2019-01-01 | Plaintiff began using its technology (approximate date) |
| 2020-03-17 | ’706 Patent Issued |
| 2021-06-15 | Plaintiff sent first notice letter to Defendant |
| 2021-07-21 | Plaintiff sent second notice letter to Defendant |
| 2021-10-01 | Defendant's agent acknowledged receipt of notice letter |
| 2022-02-09 | First Amended Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 10,592,706, "Artificially Intelligent Order Processing System," issued March 17, 2020
The Invention Explained
- Problem Addressed: The patent describes inefficiencies in traditional QSR ordering, such as employee time constraints leading to customer frustration and missed up-sell opportunities (’706 Patent, col. 2:1-15). It notes that existing automated speech systems have limitations, including small vocabularies and difficulty recognizing different accents, speeds, or tones of voice (’706 Patent, col. 2:16-28).
- The Patented Solution: The invention is a hybrid automated ordering system that combines artificial intelligence for speech recognition and order processing with a human "auditor" for oversight and intervention (’706 Patent, Abstract). The system converts a customer's speech to text, uses a natural language processor to assemble an order, and, critically, includes a mechanism for an auditor (who may be off-site) to be alerted to exceptions, review the order, and potentially deactivate or override the automated system to ensure accuracy (’706 Patent, col. 6:36-54; FIG. 4). This human-in-the-loop architecture is designed to handle errors or customer confusion that a fully automated system might miss.
- Technical Importance: This approach sought to improve the reliability and user acceptance of voice-based ordering systems by augmenting AI with human intelligence for exception handling, a key failure point for earlier automated systems (’706 Patent, col. 2:22-28).
Key Claims at a Glance
- The complaint asserts independent Claim 1 (Compl. ¶35).
- Essential elements of Claim 1, a method claim, include:
- Providing an audio stream of a customer's order.
- Using an AI-trained speech recognition module to convert the audio to text.
- Using a natural language processor to create or modify an order from the text.
- Generating an order with the natural language processor.
- Alerting an off-site auditor, connected via a data connection, of detected exceptions.
- Deactivating the order processor with the auditor in response to the exceptions.
- Prompting an on-site employee to engage the customer and complete the order.
- The complaint does not explicitly reserve the right to assert dependent claims.
III. The Accused Instrumentality
Product Identification
The accused instrumentality is Defendant's "Kea Voice" conversational AI technology (Compl. ¶14).
Functionality and Market Context
- Kea Voice is a software system designed to automate customer order-taking for restaurants, particularly over the phone (Compl. ¶¶13-14). The system is described as using "natural language processing and machine learning" to convert voice orders into digital text orders (Compl. ¶14). The complaint includes a marketing image for Kea Voice stating it "Gets smarter the more it is used" (Compl. p. 5).
- A central allegation is that the Kea Voice system uses "highly trained agents" who are "always ready to intervene when the AI lacks ability to understand the customer" (Compl. ¶23). These agents are alleged to function as the claimed "auditor," detecting errors and intervening in the order process (Compl. ¶24). The system is also alleged to integrate with restaurant Point of Sale (POS) systems to prompt on-site employees to fulfill the completed order (Compl. ¶26).
- The complaint alleges Kea Voice is commercially significant, being in use at "250 restaurants operated by seven brands," including Five Guys and Domino's franchises (Compl. ¶27).
IV. Analysis of Infringement Allegations
’706 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| providing an audio stream of an order of a customer positioned on a site | The Kea Voice system receives an audio stream from a customer giving an order over the phone (Compl. ¶17). | ¶17 | col. 10:20-24 |
| providing an order processor having a speech recognition module trained using artificial intelligence programs | Kea Voice is alleged to operate using "natural language processing and machine learning" and "artificial intelligence that 'learns to understand accents and special requests better'" (Compl. ¶¶14-15). | ¶¶14-15 | col. 3:40-45 |
| converting a word or words in the audio stream to text using the speech recognition module | Kea Voice allegedly converts spoken audio from a customer to a text order with its conversational AI software, which is alleged to contain a speech recognition module (Compl. ¶18). | ¶18 | col. 3:45-48 |
| processing the text communication... to identify a word or words in the text of the converted audio stream according to a previous spoken word training | The system is described as using "natural language processing" and creating digital textual orders from customers' spoken words (Compl. ¶¶19, 22). | ¶¶19, 22 | col. 3:48-53 |
| providing a natural language processor having order assembly capabilities and exception detection capabilities... | Kea Voice allegedly uses a natural language processor for order assembly and is alleged to have exception detection capabilities via "call center backups" and "highly trained agents" who intervene when the AI cannot understand the customer (Compl. ¶¶22-23). | ¶¶22-23 | col. 3:53-60 |
| generating an order with the natural language processor | The system is alleged to create "digital textual orders" from spoken words using its natural language processor (Compl. ¶22). | ¶22 | col. 3:60-61 |
| alerting an auditor of detected exceptions in the order, the auditor located off-site and connected to the order processor via a data connection | On information and belief, Kea Voice is alleged to use an "outside agent, or auditor, who is ready to detect errors... and intervene" and is connected via a data connection (Compl. ¶¶24-25). | ¶¶24-25 | col. 4:62-67 |
| deactivating the order processor with the auditor in response to the detected exceptions in the order | The complaint does not use the word "deactivating," but alleges that "highly trained agents are always ready to intervene," which suggests a potential override or deactivation of the automated process (Compl. ¶23). | ¶23 | col. 18:32-35 |
| prompting an on-site employee to engage the customer and complete the order | Kea Voice is alleged to provide digital orders to screens or monitors, which "prompts the on-site employees to fulfill the order" (Compl. ¶26). | ¶26 | col. 4:50-54 |
Identified Points of Contention
- Scope Questions: A central question may be whether Kea's "highly trained agents" who "intervene" (Compl. ¶23) perform the specific functions of the claimed "auditor." The analysis will depend on whether this intervention constitutes "alerting" the agent and "deactivating the order processor" as required by the claim language.
- Technical Questions: What evidence does the complaint provide that the Kea Voice system performs the step of "deactivating the order processor with the auditor"? The complaint alleges agents "intervene" (Compl. ¶24), but the factual basis for how this intervention technically occurs and whether it amounts to a deactivation of the AI processor is not detailed and is based on "information and belief."
V. Key Claim Terms for Construction
The Term: "auditor"
- Context and Importance: This term is central to the patent's asserted novelty over purely automated systems. The definition of "auditor" and its required functions will be critical to determining if Kea's "highly trained agents" (Compl. ¶23) meet this limitation. Practitioners may focus on this term because the infringement theory hinges on mapping the role of Kea's human agents to the specific role of the claimed auditor.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The patent describes the auditor's role functionally, such as to "review the order to detect errors or faults" and to "take-over the order processing system" (’706 Patent, col. 8:28-32), which could support a construction covering any human agent who performs these functions.
- Evidence for a Narrower Interpretation: The patent describes the auditor as being "physically located off-site" and having specific capabilities, such as reviewing a text summary of requests and approving or editing the NLP's response ('706 Patent, col. 6:40-54). This could support a narrower construction requiring these specific attributes.
The Term: "deactivating the order processor"
- Context and Importance: This active step is a key part of the claimed human-in-the-loop intervention process. Whether Kea's system performs this specific action, as opposed to merely flagging an order for review or supplementing the AI's output, will be a focal point of the infringement analysis.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The term could be construed broadly to mean any action where the auditor's input takes precedence over the automated processor's output for a given exception, effectively sidelining the processor for that decision.
- Evidence for a Narrower Interpretation: The specification provides a more concrete example, stating the auditor has the "capability to... switch the artificially intelligent order processing system off to pass control to on-site employee(s)" ('706 Patent, col. 6:55-58). This could support a narrow construction requiring a full transfer of control away from the automated system.
VI. Other Allegations
- Indirect Infringement: The complaint alleges inducement of infringement by asserting that Defendant instructs its restaurant customers on how to implement and use the Kea Voice technology in a way that performs the claimed method (Compl. ¶44). It also alleges contributory infringement, claiming the Kea Voice system is not a staple article of commerce and is not suitable for a substantial non-infringing use (Compl. ¶51).
- Willful Infringement: Willfulness is alleged based on Defendant’s purported actual knowledge of the ’706 Patent since at least June 15, 2021, as a result of multiple notice letters sent by Plaintiff, and its continued alleged infringement thereafter (Compl. ¶¶ 28-31, 38).
VII. Analyst’s Conclusion: Key Questions for the Case
- A central issue will be one of evidentiary proof: can Plaintiff, through discovery, demonstrate that Defendant's "highly trained agents" perform the specific, multi-step intervention recited in Claim 1? The complaint relies on inferences from marketing materials and allegations on "information and belief" to map the functions of the agents to the claimed "auditor," particularly for the "alerting" and "deactivating" steps.
- The case may also turn on a question of claim scope: can the phrase "deactivating the order processor," in the context of the patent's specification, be construed to cover the "intervention" allegedly performed by Defendant's agents? The court's interpretation of this term—whether it requires a complete system hand-off or allows for a more limited override—will be critical to the infringement analysis.