SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activity:

Research into the use of the RAG method in the study of organisations and procurements.

The following research results and deliverables have been achieved for the activity:

As a result of the activity, a prototype that uses the RAG method was developed under laboratory conditions. It collects up-to-date information on the requirements that apply to procurements and organisations:

• Comparison of retrieval strategies: fragment vector search was compared with an agentic approach, in which the AI model reads the procurement documents on demand. The agentic approach was implemented in the prototype.

• Document structure recovery component: recognition of PDF paragraphs and headings, processing of archive contents, and handling of diacritical marks in search.

• Retrieval module for the list of documents to be submitted: a specialised AI agent scans the procurement regulations across five topics (submission, format and language, qualification, finance, technical part). For each document, it states the category, whether it is mandatory, and a verbatim quote from the source. More than 100 procurements were processed.

• Deterministic citation verification: each citation is automatically compared with the source text and shown to the user with a colour signal (confirmed / partial / not found). This lets the user check where each AI entry comes from.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”.

Over the past 3 months (1 April 2026 to 30 June 2026), SIA Sorsera has carried out and completed the following activity:

Research into the core functionality of an agentic communication solution.

The following research results and deliverables have been achieved for the activity:

Research into the core functionality of an agentic communication solution was carried out in public procurement communication scenarios:

• A set of communication context retrieval tools was developed: search for suppliers and procurements, access to procurement documentation, collection of market information, web search and research of legal information.

• A multi-agent delegation model was validated, in which the main conversation agent delegates tasks to specialised sub-agents.

• Three communication automation use cases were prototyped: drafting messages, suggesting questions about procurement procedures, and suggesting replies to questions from the contracting authority, with preliminary analysis of the procurement documentation.

• Agentic interaction was implemented in the platform interface and validated with published procurements.

SIA Sorsera turpina realizēt pētniecības projektu “Automatizēta piegādātāju monitoringa un analīzes risinājuma prototipa izveide”.

SIA Sorsera pēdējo 3 mēnešu laikā (01.04.26-30.06.26) veikusi un noslēgusi sekojošu aktivitāti:

Aģentiskas saziņas risinājuma pamatfunkcionalitātes izpēte.

Ir sasniegti aktivitātei atbilstošie pētniecības rezultāti un nodevumi:

Veikta aģentiskas saziņas risinājuma pamatfunkcionalitātes izpēte publisko iepirkumu saziņas scenārijos:

• Izstrādāta saziņas konteksta izguves rīku kopa: piegādātāju un iepirkumu meklēšana, iepirkuma dokumentācijas piekļuve, tirgus informācijas apkopošana, tīmekļa meklēšana un juridiskās informācijas izpēte.

• Aprobēts daudzaģentu deleģēšanas modelis, kurā galvenais sarunas aģents deleģē uzdevumus specializētiem apakšaģentiem.

• Prototipēti trīs saziņas automatizācijas lietošanas gadījumi: ziņojumu melnrakstu sagatavošana, jautājumu ieteikšana par iepirkuma procedūrām un atbilžu ieteikšana uz pasūtītāja jautājumiem ar iepirkuma dokumentācijas priekšanalīzi.

• Aģentiskā mijiedarbība ieviesta platformas saskarnē un validēta ar publicētiem iepirkumiem.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

The following research results and deliverables have been achieved for the activity:

• Development of a conversational agent embedded in the procurement view. It lets users ask questions about a procurement and its documents in natural language, and gives source-based answers with references;

• Creation of document ingestion and indexing (text extraction, chunking, vector embedding, KNN index) with asynchronous processing, status tracking and resilience mechanisms;

• Implementation of semantic search and automatic extraction of structured data (procurement deadlines, requirements, submission format);

• Development of an agent toolset (6 tools: document search, record keeping, summaries, finding similar procurements, terminology, web search) and a multi-provider AI architecture with unified cost tracking.

1 January 2026 - 31 March 2026: SIA Sorsera continues to carry out research project No. <DI-10> “Development and Validation of Core Functionality of a Supplier Analysis Solution”.

The following research results and deliverables have been achieved for the activity:

• A component was created: document-centric RFx management

• A unified document metadata schema for documents created in the platform and documents obtained externally

• An OpenSearch indexing layer with unified search across both document types

• Event-driven architecture (AWS EventBridge), prepared for agent integration

• Multi-tenant data isolation and permission model

• Knowledge gained about the document-centric approach

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• “Use of AI algorithms to improve the recommendation system”

The following research results and deliverables have been achieved for the activity:

• Development of a semantic search engine that lets users search for suppliers and organisations with natural language queries, not only exact keywords;

• Integration of vector embedding models, testing models of different dimensions (1536 and 3072 dimensions) and evaluating how well they suit different search scenarios;

• Creation of a hybrid search architecture that combines semantic (vector) search and traditional keyword search, which gives higher result accuracy;

• Development of an agentic search component in which an AI algorithm interprets the user’s natural language question, automatically generates structured search filters and gives AI-based explanations of how well each result matches.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”

The following research results and deliverables have been achieved for the activity:

• “Org Merge” tool with an 80%+ reduction in duplicates

• Organisation monitoring profile functionality

• Automated email notifications about procurement activity

• Configurable monitoring scenarios (buyers, competitors, industries)

• Validation with beta users

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• Use of generative AI algorithms for document creation.

The following research results and deliverables have been achieved for the activity:

• As a result of the activity, a functional AI-based document generation prototype was developed under laboratory conditions:

1. Technical requirements generation module - the user enters a description of the procurement object, and the system automatically generates structured technical specifications, including functional requirements, quality standards, delivery conditions and compliance criteria.

2. Multi-layer AI architecture - the prototype calls three different AI models in parallel, which gives a variety of results and the option to choose the best variant. It is also possible to replace both the generating model and the evaluation model (including with other LLM API providers, e.g. Claude).

3. Automated quality assessment system - an AI-based comparison tool that assesses the quality of the generated documents against five criteria that matter in EU procurement practice, with numerical scores and visualisation.

• Quality assessment reports were prepared on how well the generated RFx documents meet the requirements of EU public procurement.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”

The following research results and deliverables have been achieved for the activities:

• Within the activity, we developed the core functionality of the supplier analysis solution. We built a system that automatically collects and processes data from various procurement databases (TED, EIS, IUB, Riigihanked, VPT, CERN, Mercell and other private sources), company registers and public information. We developed a data processing pipeline that standardises and enriches information from different sources.

• We integrated OpenAI text-embedding models (large and small) for supplier data analysis. The system can automatically explain why a given supplier matches the search criteria. We built functionality for analysing the procurement history of suppliers: contracts won, largest clients and participation in tenders.

• We added a translation step (Google Translate) and information enrichment to the data processing pipeline. As a pilot project, we built a prototype for an international scientific organisation, integrating ~12,000 supplier records with semantic search. We validated the solution with potential users, who confirmed the need for the functionality.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• Use of AI algorithms for data interpretation.

The following research results and deliverables have been achieved for the activities:

- As a result of the activity, AI-based data interpretation components were successfully integrated into the Sorsera platform:

1) The summary generation function produces structured and informative summaries of procurement specifications, making complex technical documentation easy to understand;

2) The keyword extraction function automatically identifies critical terms and requirements in procurement documentation in Latvian, significantly reducing the time needed for manual analysis;

3) The organisation profile analysis tool gives in-depth insight into the business areas, experience and capacity of companies, based on historical data and public information;

- The interactive system will let users ask questions freely in natural language and receive contextually accurate answers;

- Quality assurance mechanisms were put in place in line with industry best practice: validation layers to reduce hallucinations and repeated verification of outputs.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”

The following research results and deliverables have been achieved for the activities:

- A detailed report, “EU Supplier Data Sources for Automated Monitoring”, was created. It identifies data sources that are ready for integration into the Sorsera automated supplier monitoring solution, and summarises the knowledge gained on:
1) Financial data infrastructure: Free and low-cost sources for monitoring the financial health of companies were identified, including the GLEIF API with daily updates and a connection to the EU insolvency register.
2) Procurement performance tracking: Access was mapped to the TED database with 740,000+ procurements a year, to national portals and to the PEPPOL network with 1.4 million participants.
3) News and event monitoring: Methods for automated tracking of company news from multiple sources were tested.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 3 months, SIA Sorsera has carried out and completed the following activities:

• Use of AI algorithms for data search.

The following research results and deliverables have been achieved for the activities:

• As a result of the activity, AI classifiers were integrated into the Sorsera platform, which significantly improves how data is organised and how easily it can be found;

• The classifiers split procurements reliably across 50 industries, which makes the data easier for users to understand and use for market research;

• Organisations were classified by industry, which makes it faster to select and compare them;

• Classification accuracy was assessed with metrics that exceeded the initial expectations in several tests;

• The ability of GPT models to work in both Latvian and English was demonstrated, which significantly widens the use of the Sorsera platform in the EU.

SIA Sorsera is starting the research project “Development of a Prototype of an Automated Supplier Monitoring and Analysis Solution”.

Total implementation period: from 1 April 2025 to 30 June 2026.

Total eligible project costs: EUR 396,000.00, including a grant of EUR 296,800.00.

The project is carried out with co-funding from European Union funds, under Cabinet of Ministers Regulation No. 34 of 18 January 2024 (regulations for the implementation of investment 2.2.1.3.i “Support for the introduction of new products and services in business”, under reform and investment direction 2.2 “Digital transformation and innovation of enterprises” of the Latvian Recovery and Resilience Facility Plan), within the framework of a competence centre.

SIA Sorsera continues to carry out the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:

• Preparation and initial validation of the research environment for AI algorithms.

• Use of AI algorithms for data search.

The following research results and interim deliverables have been achieved for the activities:

• An environment was created for systematic collection and processing of data from various data sources, and for passing that data to AI analysis.

• In-depth research was carried out into leading AI processing services, including the use of APIs and practical validation in the area of data search.

• Several POC (proof of concept) tests were carried out with positive results.

• The ability of the Perplexity API to accurately identify and retrieve organisation and procurement data from our existing database was successfully demonstrated.

• A locally run component was developed that demonstrates the functionality of the search algorithms, including on publicly available and collected data about organisations.

• A significant increase in data search efficiency was found compared with traditional data retrieval methods (a report was created on the quantity and quality of the data obtained).

SIA Sorsera has completed the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

The following results and deliverables have been achieved for activity 8 and interim stage 4:

• Summary and conclusions of user tests.

• Screenshots of the digitalisation solution deployed in a real environment.

• Project summary documentation.

All planned project results have been achieved, including:

• A functional prototype of the requirements definition tool was created.

• A digital library prototype was created.

• The prototypes were deployed and successfully tested in a real industry environment.

• Technology readiness level TRL 7 was reached.

The project results have also been published in 2 scientific publications in a SCOPUS-indexed journal.

SIA Sorsera is starting the research project “Development of a Prototype of an Artificial Intelligence-Based Supply Chain Solution”.

Total implementation period: from 1 September 2024 to 28 February 2027.

Total eligible project costs: EUR 1,271,880.00, including Recovery and Resilience Facility co-funding of EUR 952,410.00.

The project is carried out with co-funding from European Union funds, under Cabinet of Ministers Regulation No. 34 of 9 January 2024 (regulations for the implementation of investment 2.2.1.3.i “Support for the introduction of new products and services in business”, under reform and investment direction 2.2 “Digital transformation and innovation of enterprises” of the Latvian Recovery and Resilience Facility Plan), within the framework of competence centres.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

The following research results and deliverables have been achieved for the activity:

• Research results of the activity were achieved

• Summary and conclusions of the core functionality tests of the overall digitalisation solution

• Screenshots of the digitalisation solution

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:

Development of the core functionality of the requirements definition tool. Development of the core functionality of the digital library software. Testing and validation of the overall digitalisation solution (reaching TRL 5). Creation of the prototype development plan. The research results and interim deliverables for the activities have been achieved.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:

Development of the core functionality of the requirements definition tool. Development of the core functionality of the digital library software. Testing and validation of the overall digitalisation solution (reaching TRL 5). Creation of the prototype development plan. The research results and interim deliverables for the activities have been achieved.

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:

Development of the core functionality of the requirements definition tool. Development of the core functionality of the digital library software. Testing and validation of the overall digitalisation solution (reaching TRL 5). Creation of the prototype development plan. The research results and interim deliverables for the activities have been achieved.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions” and has successfully completed activity 5 of the project:

- Development of core functionality of the requirements definition tool. The following research results and deliverables have been achieved for the activity:

• Research results of the activity were achieved

• Screenshots from the technology demonstration (requirements definition tool)

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:

• Study of the impact of the unified URDx format on procurement results.

• Research into wireframes for digitalisation solutions

• Definition of the unified URDx format

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

- A study was carried out on the impact of the URDx format on procurement results, by implementing a unified format in the procurement discussion and consultation process. Tests were carried out that let contracting authorities define requirements and let suppliers give feedback using the unified format.
- The unified URDx format, its application and the issues identified in connection with its integration were defined
- Functional solution wireframes were created for placing the URDx format on the platform.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions” and has successfully completed activity 3 of the project:

• Study of the impact of the unified URDx format on procurement results

• Research into wireframes for digitalisation solutions

The following research results and deliverables have been achieved for the activity:
- Use-case scenarios were drawn up for applying the URDx format in the procurement planning process, to maximise the positive impact of the format on procurement results.
- Wireframes were created for the digitalisation solution

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

Over the past 6 months, SIA Sorsera has carried out and completed the following activities:
- Research into the methodology of the digitalisation process, definition of requirements for the digitalisation solution
- Research into the requirements of the unified URDx format

The following research results and interim deliverables have been achieved for the activities:
- Detailed requirements for the digital tool were defined (5 requirement categories)
- Technical documentation for the URDx format was created

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

Over the past three months, SIA Sorsera has carried out the following activities:

• Research into the methodology of the digitalisation process, definition of requirements for the digitalisation solution (activity completed)

• Research into the requirements of the unified URDx format (activity started)

The research results for the activities have been achieved.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera is carrying out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”. The project is carried out under the EU support programme “Support Instrument for the Development of Innovation Clusters” from 1 November 2022 to 30 October 2024. The total project costs are EUR 545,956.00, including a grant of EUR 399,480.00.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.

SIA Sorsera completes interim stage 3 of the EU co-funded research project.

SIA Sorsera continues to carry out the research project “Prototype Development for a Real-Time Requirements Definition Tool with Digital Library Technology Solutions”.