Generative Artificial Intelligence (Gen AI) is disrupting the corporate purchasing sector, offering innovative, high-performance solutions for optimizing processes and data processing. With advanced tools like Ivalua’s Intelligent Virtual Assistant (IVA), the world of Purchasing and Supply Chain can now improve the efficiency and accuracy of their operations, while guaranteeing data security and confidentiality.
In this article, we explore the benefits of Gen AI, the existing and upcoming use cases present in IVA, and how Cyrias can support you in integrating these revolutionary technologies within your Ivalua platform. This article follows on from Cyrias Connect, an event organized with our customers, and brings a summary of the keynote while highlighting the key elements of this event.
Vocabulary
- LLM: Largue Language Model. Is a type of artificial intelligence designed to understand, generate and manipulate human language to generate content.
- RAG: Retrieval Augmented Generation. Powerful tool for combining LLM capabilities with an extensive database (document database, database, etc.).
- Prompt: instruction and context given to the LLM to guide its responses.
Introduction
GEN AI is a type of artificial intelligence, based on LLM (Large Language Model), designed to understand, generate and manipulate human language. Our first contact with this technology was in our private sphere, through ChatGPT. The technology we find today in our purchasing information systems is largely based on this architecture.
The stakes are high for today’s purchasers, whose scope of action is evolving without having more resources. If properly mastered and implemented, generative artificial intelligence can enable them to optimize non-value-adding tasks and focus once again on their core business.
To meet these challenges, Ivalua’s Virtual Assistant (IVA) offers artificial intelligence integrated into all its modules and accessible throughout the application with the same level of power. Available from version 174 of the application, more than ten turnkey plug-ins can be easily downloaded from the store and used to make the most of applications, such as the legal assistant which helps generate legal clauses, the contractual assistant which can summarize and highlight the essential points of a contract submitted to it in PDF format, or the questionnaire creation assistant which, based on a simple user prompt, can create a questionnaire.
Although it is essential to check the content that is generated, we would point out here that IVA is your assistant and not the other way around, so how does it know how to create such high-quality information?
It’s also legitimate for purchasing departments to wonder about the security of their data, which is transmitted to these models. So how can Ivalua prevent this data from being reused for model training purposes, or simply by other customers?
However, subcontracting part of the work to a virtual assistant by entrusting it with its own data can raise fears.
Use cases (technical), data quality and IVA
Featuring a proprietary architecture, IVA orchestrates user requests by enriching them with internal data and, if necessary, calls on the Internet for more recent information. Ivalua has ensured that your data cannot be used to train models, and cannot be reused by other customers, by deleting your data once the session or conversation with the assistant is over.
By adopting IVA, you can not only supercharge your purchasing processes, but also gain in efficiency and responsiveness, while preserving the confidentiality of your sensitive information.
To date, IVA offers several use cases for optimizing procurement processes. All these use cases are conceivable; it’s up to you to write your own based on your business issues, and make the most of the power of Gen AI.
- Supplier research: Gather information on suppliers.
- Legal assistant: Creation of contract summaries and drafting of clauses.
- Category & Market Insights: Research and synthesis of trends.
- Enhanced Communication: Create and enhance communications content.
- Generate Improvement Plans: Analysis and suggestions for improving supplier performance.
- Chat with your documents: Extract relevant information from large documents.
Technology’s future
Data quality:
- Enrich prompts with current data thanks to the internet search option (can be activated on the fly)
- Enhance the RAG with data from the document database (e.g. a company’s purchasing policy)
- Text-to-SQL: Ivalua is working on describing its data model so that it can be understood by an LLM, and so that IVA can generate a SQL query for us from natural language.
LLMs:
- Ivalua has built its architecture to be LLM agnostic. In this way, we can make the most of the IVA assistant by choosing the model to use according to usage. We can also imagine “plugging in” a customer’s private LLM to IVA for enhanced response quality.
- Become “multi-modal” by analyzing photos (IVA Vision) and transcribing (IVA Voice) a request made orally by the user.
Multi assistant:
- Today, IVA is a single, multi-tasking assistant. Tomorrow, Ivalua intends to integrate multiple agents with their own specificities (legal assistant, accounting assistant, SQL assistant, etc.).
- Programmatic skills: offer more advanced AI integration without necessarily having to go through the application design, for example to propose a replacement item to the user when he or she has proposed an obsolete item without having to call on the AI.
Cyrias support
At Cyrias, we are ready to support our customers in implementing Gen AI within their Ivalua Purchasing Information System. Thanks to our in-house expertise and participation in the Gen AI Bootcamp and Ivalua’s AI-focused Innovation Cup, we have acquired the skills needed to implement and configure use cases specific to your needs.
Our team is ready to help you get the most out of IVA, the tool developed by Ivalua.
Please contact us to find out more about how we can help you transform your purchasing management with the power of Gen AI and Ivalua’s advanced tools.
Creation by 