We are proud to announce that OPTI Software is among the finalists of the 2026 edition of the IT&C Industry Excellence Awards, organized by ANIS (the Romanian Software and Services Industry Association).
Our applied research project, "AI Architecture for B2B in 2026", has been shortlisted in the prestigious category R&D Program of the Year. Winners will be announced at the ANIS Gala taking place next Tuesday, on 12.05.2026, starting at 18:30 (ANIS website).
A competition at the highest level
This year's edition confirmed the high standards of the industry, with a strong emphasis on concrete projects. The fact that our innovation reached the final is all the more honorable given that the same category featured research and development projects from tech giants such as Adobe Systems Romania, Cognizant Technology Solutions and DB Global Technology.
We are delighted to be on the finalists' podium alongside BRINEL | iQanto and Tquila Automation.
About the nominated project: Engineering, not magic
The project that convinced the ANIS jury addresses a systemic problem in the real economy: the blockage of distribution and manufacturing companies locked in legacy ERP systems, and the inherent hallucination risk of AI models.
Our R&D effort resulted in a hybrid architectural framework that makes AI safe for Enterprise and B2B environments (pharmaceutical, distribution, industry). In short, we reversed the classic paradigm: business rules dictate, and AI executes.
Through applied technical innovations such as Safety Guardrails, Data Contracts and a Privacy-First Middleware, the framework developed by OPTI ensures that AI algorithms (Vector Search, GraphRAG) only recommend products that are in stock, respect minimum profit margins and are 100% technically compatible. This transforms passive ERPs into active sales engines.
"In B2B, AI does not work as a form of magic, but as precision engineering. It must follow the company's business rules, not invent them."
Marian Calborean, Founder of OPTI Software
Research on AI across 6 domains
The OPTI team segmented the AI challenge for industry into six domains, with plans to detail each at the highest standards, backed by international quantitative reports.
Why do companies want AI in 2026? Three identified opportunities:
1. ERP as a sales engine: Growing sales and generating recommendations based on transaction history.
Why? Studies cite an average 15% increase in ecommerce orders with personalization and AI. Also, a 20% performance boost for junior sales agents assisted by AI.
2. Multimodal input: Automating order intake from images, voice and PDFs, including WhatsApp or public procurement platforms.
Why? Text and voice recognition technologies for the Romanian language now achieve over 90% accuracy.
3. Auto-correction: Learning from customer behavior and forecasting required stock levels.
Why? Organizations with adaptive AI scale internal projects 2.5 times faster, and widely available AI models can be retrained daily, giving companies greater adaptability.
What are companies' concerns and the AI architecture solutions?
4. The search paradox (keywords vs. intent)
Eliminating search errors (colloquial jargon vs. correct product names) through algorithms that understand technical shorthand while simultaneously respecting exact product codes ("SKU-8521" vs. "SKU-8522").
Why? Search is critical in retail: 43% of ecommerce visitors go directly to search. In the B2B supply chain, 75% of operators say they would switch suppliers for a better online experience.
Hybrid search (classic lexical plus AI semantic) is the architecture detailed as the solution.
5. Data integrity: Minimizing AI hallucinations and unifying company data into a single source of truth.
Why? An employee loses an average of 9 hours per week searching for information across different systems, and AI-generated false data is not acceptable in B2B.
Using classical databases in parallel for sensitive details such as price and stock is the architecture detailed as the solution.
6. Security: Compliance with the NIS2 directive, GDPR and the ISO 27001 standard, and protection of trade secrets.
Why? According to cited international research, 70% of innovative companies protect their trade secrets and 78% of IT managers require transparency and auditability in AI.
AI technologies widely available today allow data storage strictly within the EU, prevent the model from learning from company prompts, and enable automated checks on legacy software, all of which can significantly strengthen a company's security posture.
Results and open-source guide
The results of this research are already applied in our SaaS platform (AI Sales), which significantly reduces the workload of sales agents in technical sectors.
The results will also be available open-source as a series of 6 technical guides currently being published, with over 100 technical diagrams in total, designed to set a new industry standard.
The first Guide was officially launched at the BeeGlobal event on 10.03.2026 (see details):
Read more
Discover the nominated project: AI Architecture for B2B 2026
See the AI Sales automation software
Download Guide #1: Smart recommendations, upsell and rules
The OPTI Software team is looking forward to next Tuesday's event! We wish all finalists innovating in Romanian IT the best of luck.
