Leveraging the Qthena Edge: Legal Analysis

16 February 2024

In the rapidly evolving landscape of legal technology, the quest for efficiency and accuracy is paramount. At the heart of this pursuit lies the challenge of integrating sophisticated AI tools into the legal workflow without the cumbersome need for prompt engineering. Qthena by ipQuants AG stands out as a pioneering solution, transforming how legal professionals interact with generative AI.

The Technical Challenge Unveiled

Our recent comparative study put Qthena and ChatGPT4 to the test, presenting both platforms with a complex legal question. This challenge required not only a deep understanding of legal and technical information but also the ability to craft strategic legal arguments—traditionally the domain of experienced patent attorneys.

We provided ChatGPT by OpenAI and Qthena with the same four PDF documents – Search opinion, the application as filed, and prior art documents D1 and D2 – and asked the same question as follows:

Question asked:  Please based on the NEW_CLAIM_1 prepare an inventive step argument, where prior art D1 is the closest prior art. Argue why the skilled person starting from prior art D1 would not arrive at the subject matter of revised claim 1, even if the skilled person would consider prior art D2 in combination with prior art D1.

The NEW_ClAIM_1 is: 1. A solid state power controller, comprising a power supply line (110) configured to be connected between a power source (200) and a load (300); a semiconductor switching unit (120) provided on the power supply line (110) and configured to switch between at least three states, including a conducting state, a non-conducting state and an intermediate state, according to a command signal, wherein the semiconductor switching unit (120) is configured to, in the intermediate state, limit a current on the power supply line (110), according to the command signal, to be below a predetermined current threshold; and wherein the semiconductor switching unit (120) is formed by one or more non-programmable electronic components; a state machine (130) configured to exhibit at least three states including an ON state, an OFF state and a pre-charge state; output the command signal according to a current state of the state machine (130); the command signal indicates the semiconductor switching unit (120) to be in the intermediate state if the current state of the state machine (130) is the pre-charge state; characterized in that the state machine (130) is a non-programmable logic part formed by a plurality of non-programmable electronic components and is configured to allow for transitions from the OFF state to the pre-charge state, from the pre-charge state to the OFF state, from the pre-charge state to the ON state, and from the ON state to the OFF state.

ChatGPT4 vs. Qthena: A Clear Distinction

The results were telling (see full answers below). ChatGPT4, while capable, fell short of delivering the attorney-level analysis expected in legal argumentation. Its response lacked the nuanced understanding and structured reasoning that come second nature to legal professionals.

For example, a seasoned European Patent Attorney would have understood that the question posed requires reasoning along the so called problem-solution approach as defined in the Guidelines for Examination at the European Patent Office. This approach, requires justification why the applicant selects a particular document as the “closest prior art”, what the “objective technical problem” is, and why the skilled person would not arrive at the claimed subject matter (so called could-would approach).

Qthena’s response (see below) demonstrated a profound grasp of legal intricacies, delivering a well-structured and strategic legal argument. Qthena’s answer encompassed the precise terminology and legal structure expected for such an argument analysis, showcasing its unparalleled ability to provide expert-level insights. The results, delivered in a few seconds, are at a quality level corresponding to a senior european patent attorney.

Despite the user not specifying the problem solved by the invention or any “Objective Technical Problem” (see the above “Question”) , Qthena, leveraging superior ipQuants prompt engineering and data, intuitively identifies and includes such critical details for inventive step argumentation. It also autonomously incorporates the “Could-Would Approach” when necessary. This ensures the delivery of comprehensive legal arguments, maintaining the highest quality even if the user omits specific instructions.

Why Qthena Stands Apart

  • Out-of-the-Box Prompt Engineering: Qthena revolutionizes prompt engineering by rendering user input unnecessary in this domain. With ipQuants’ deep integration of GPT4, the platform instinctively understands and adapts to your legal workflow right from the start. This enables you to focus entirely on your tasks, achieving desired outcomes through straightforward interactions with Qthena. Just pose your questions clearly, as you would naturally, and let Qthena handle the rest.
  • Workflow Integration: Qthena is uniquely designed as the ultimate co-pilot, featuring a dedicated cockpit for unparalleled workflow integration. It provides instant access to all necessary data and information, equipping it with the expertise of a senior patent attorney within seconds. This integration not only significantly cuts down on startup time but also ensures that users save considerable time, making Qthena a game-changer in legal tech efficiency.
  • Time and Resource Savings: By streamlining legal processes, Qthena offers substantial time and resource savings. Its efficiency allows legal professionals to accomplish tasks more swiftly and with greater precision, allocating more time for direct client engagement. Qthena is not just a tool but a transformation in how legal services are delivered, optimizing productivity without compromising quality.

Embracing the Future of Legal Analysis

Qthena represents the epitome of customer obsession in legal tech. By offering a solution that integrates seamlessly into existing workflows, Qthena sets a new standard for legal analysis tools. Our case study illustrates not just the potential of AI in law but the realization of that potential through deep workflow integration.


Qthena’s superiority over ChatGPT-4 isn’t surprising, as OpenAI’s CEO, Sam Altman, notes the full potential of large language models (LLMs) like GPT-4 is realized when they’re integrated into specific workflows. Qthena excels because it’s designed for this, using proprietary ipQuants data and prompt engineering for seamless workflow integration. Users interact with Qthena as effortlessly as with any team member, highlighting its tailored utility.

Uncompromised Privacy and Cutting-Edge Security with Qthena

In today’s digital landscape, data security and privacy are paramount (we wrote in detail about this aspect in this previous blog post; Open Article). At ipQuants AG, we understand the criticality of protecting your sensitive information. That’s why Qthena is built on the foundation of state-of-the-art Azure Datacenters, ensuring the highest standards of data security and privacy. Recognizing the diverse needs of our global users, we offer flexible hosting solutions. Whether you require hosting entirely within the European region or in other specific areas to meet regulatory requirements, Qthena adapts to your organisational needs.

Qthena answer:

*** Start of Inventive Step Argument

– Inventive Step when starting from D1 as the closest prior art

The subject matter of new claim 1 fulfills the requirements of an inventive step for the following reasons:

– Closest Prior Art D1

D1 discloses a solid state power controller with a controller that limits power dissipation based on measured voltage across the solid state switch. Although D1 addresses the control of power dissipation, it does not disclose the specific use of a non-programmable state machine with distinct states for controlling a semiconductor switching unit.

– Objective Technical Problem

The objective technical problem can be formulated as providing a solid state power controller with a reliable and deterministic state transition mechanism that minimizes complexity, enhances safety, and simplifies the certification process for aviation technology applications.

– Could-Would Approach

Even if the skilled person starting from D1 would consider the teaching of prior art D2, which discloses an intelligent solid-state circuit breaker controlled by a digital signal processor, they would not arrive at the subject matter of new claim 1. D2 teaches an approach that relies on digital processing and programmability to control the operation states of the circuit breaker, which is fundamentally different from the non-programmable, hardware-based logic control required by the claimed invention.

D1, being focused on power dissipation control, does not provide any guidance towards implementing a non-programmable state machine nor suggests the concept of an intermediate state for inrush current management. The combination of D1 and D2 would not lead the skilled person to the invention as claimed in NEW_CLAIM_1 because the core principle of the claimed invention—non-programmable control for critical functionalities—is not hinted at or suggested by either D1 or D2. The novel concept of a non-programmable state machine with specific state transitions to manage inrush current through an intermediate state remains untaught and unobvious.

– Conclusion

For the above reasons, the subject matter of new claim 1 fulfills the requirements of an inventive step. The non-programmable nature of the state machine and its capability to control the semiconductor switching unit through a unique intermediate state for current limitation represent an inventive contribution over the combined teachings of D1 and D2. The skilled person, even when considering the teachings of D2, would not be led to the claimed invention starting from D1.

*** End of Inventive Step Argument

This inventive step argument is structured around the problem-solution approach, emphasizing that the claimed SSPC’s non-programmable state machine and the intermediate state for inrush current management would not be obvious to a skilled person familiar with D1 and D2. The argument is based on the fact that neither D1 nor D2, alone or in combination, provide any indication or motivation to develop a SSPC with the specific features claimed in NEW_CLAIM_1.

ChatGPT4 answer:

Prior art D1 discloses a solid state power controller (SSPC) designed to limit inrush current by controlling the power dissipated in the solid state switch based on a measured voltage across the solid state switch and a predetermined power dissipation threshold. This system employs a controller to adjust the output current or voltage control signal of the solid state switch to ensure the actual power dissipation does not exceed this threshold. The SSPC uses at least one solid state switch and includes features for remote control and protection, with the capability to operate the solid state switch in a linear mode to control its equivalent resistance. The document also outlines methods for implementing power dissipation regulation through various means, including a look-up table and proportional-integral control, to manage the inrush current and power dissipation effectively.

Comparatively, NEW_CLAIM_1 introduces a solid state power controller with unique features that distinguish it from D1, including:

– A specific configuration involving a power supply line, a semiconductor switching unit capable of operating in at least three states (conducting, non-conducting, and an intermediate state to limit current below a predetermined threshold), and a state machine exhibiting multiple states with defined transitions.

– The semiconductor switching unit is formed by non-programmable electronic components, and the state machine is also a non-programmable logic part formed by a plurality of non-programmable electronic components, emphasizing a distinct implementation approach compared to D1’s focus on control strategies for existing solid state switch technologies.

The inventive step argument centers on the fact that, while D1 provides a foundation for controlling inrush current through power dissipation management in a solid state switch, NEW_CLAIM_1 advances the technology by integrating a semiconductor switching unit with non-programmable components capable of a unique intermediate state for current limitation, combined with a non-programmable state machine enabling specific state transitions. This novel combination of features for managing power distribution and limiting inrush current, with a detailed emphasis on non-programmable components and the functionality of the semiconductor switching unit, represents a significant technological advancement over D1’s disclosure. The specifics of NEW_CLAIM_1, particularly the detailed structure and operational capabilities of the semiconductor switching unit and the state machine, would not be obvious to a skilled person based on D1 alone or even when considering D1 in combination with other prior art like D2, due to the distinct approach towards achieving controlled power distribution and current limitation.

Follow ipQuants on LinkedIn and sign up to our Newsletter here.

About ipQuants AG: ipQuants AG is a premier Software-as-a-Service (SaaS) company based in Schaffhausen, Switzerland. At the forefront of innovation, we offer Qthena, the only generative AI-powered co-pilot equipped with a purpose-built digital cockpit. Qthena sets a new standard for efficiency and strategic analysis for knowledge workers worldwide by unifying actionable insights with workflow automation.

For more information, please visit www.ipQuants.com