SME-Lender_3200x1200_Hero-bg

AI-powered automation transforms SME lending into a more efficient and scalable operation

65%

reduction in loan applications processing time

85%

lower cost per application processing

3x

more loan applications processed

The Impact

An SME lender with a portfolio of over $1B in SME loans partnered with Zitec to transform its loan application process through AI-powered automation.

The collaboration addressed the challenges of manual data entry, lengthy processing times and the need for more efficient risk assessment, while ensuring alignment with strict security standards imposed by regulators. By implementing a multi-step tailored AI solution, the company has tripled its processing capacity without expanding headcount, while accelerating response times and improving customer satisfaction.

This transformation has turned our partner into a resilient, AI-ready organization prepared to face unexpected future challenges, enabling them to scale easier, with streamlined operations, all while offering faster, more reliable services to its SME clients, contributing to their growth and success. The AI-driven platform automates the initial intake of loan applications and further enhances data analysis, enabling faster, better informed lending decisions.

“Partnering with Zitec to build this AI-powered solution has transformed our loan application process and tripled our operational capacity basically overnight. This has been a game-changer for our scale. We’ve moved away from manual bottlenecks and are now faster, and more accurate in our data analysis. This truly helps with our process of becoming a truly resilient AI-ready organization.”

Vice-President

The Context

Small and Medium Enterprises (SMEs) are the backbone of economies worldwide, driving innovation, employment and economic growth. A Fundera report shows that only about 48% of small businesses meet their financing needs, with 20% getting loans and 28% lacking enough capital without a loan1. That means a shocking 52% of all businesses get no financing or only get a portion of what they need, impacting markets in multiple ways.

Small businesses and their financing needs

  • 20%Get loans
  • 28%Lack enough capital without a loan
  • 52%Get no financing, or only a portion of what they need

Source: Fundera

Traditional lending processes are often delayed by manual data entry, lengthy processing times, and inefficient risk assessments, leading to delays and potential lost opportunities. These inefficiencies are exacerbated by outdated methods of data collection or a lack of customized loan products for small businesses.

However, digital transformation processes have the power to address these issues through automation and AI. By streamlining data collection, enabling faster approvals, and offering more tailored financing options, technology is helping to reduce the barriers small businesses face in securing the capital they need.

The company, a tech-enabled, forward-thinking lender, recognized these challenges and sought to transform the loan application process for SMEs. By partnering with Zitec, they aimed to leverage AI-powered automation to streamline operations, reduce operational costs and increase customer satisfaction.

Meet our Financial Services leadership team

Results & tech excellency in one package

Lucian Daia

Financial Services Solutions Director & CTO

Iulian Oprea

Tech Consulting Director

The Challenges

According to an Accenture report, financial institutions implementing Generative AI achieved, on average, a 11% reduction on operational costs within the first 18 months2. This highlights the growing need for innovative solutions that can improve operational efficiency in the lending sector. Moreover, a quick response time to loan applications is key to improving the overall customer experience.

Our client faced several key challenges before partnering with Zitec for this new project:

  • A significant amount of time and resources were spent on manually processing data from thousands of loan applications (with hundreds of thousands of pages of documents) monthly, leading to inefficiencies and delays.

  • Traditional loan processing methods put a strain on response times, impacting customer satisfaction and potentially leading to lost business.

  • The team was looking to maintain their accurate and efficient way to analyze client data (both company and owners data) and assess risk, while significantly scaling their operations to meet increasing demand, without adding significant overhead.

  • High operational costs associated with manual processing and third-party tools (advanced OCR technology for financial documents processing) were impacting business margins.

“It was time to significantly improve this process, and Zitec were our partners of choice for the development of an AI-Infused SME Lending Origination Accelerator that completely transformed our internal workflows.”

Vice-President

The Solution

In order to improve the efficiency of their operations, expand their processing capacity without the need for headcount increase and improve response time to customers’ requests, the company tasked Zitec with engineering a high-impact solution.

“Speedy negative responses that quickly eliminate potential clients who are not eligible for certain lending products, help free up the team to focus on eligible leads, ensuring that resources are allocated efficiently. This approach prevents SMEs from investing time in unfruitful applications, allowing them to focus on other opportunities. On the other hand, fast positive responses for financing enable eligible SMEs to be more agile, securing the necessary capital quickly, in order to act on growth opportunities and adapt to changing market conditions.”

Lucian Daia

CTO, Zitec

As their strategic partner, we designed the architecture and operating model from the ground up, defining the business logic and mapping all the needed workflows to ensure a seamless transition to an automated ecosystem. The result was a multi-stage AI-enabled project meant to completely transform the organization from within - the SME Lending Origination Accelerator.

The initial phase focused on operational efficiency. We developed an Azure-based AI solution that uses Azure OpenAI and Azure Document Intelligence to automate the entire data intake process by extracting and interpreting data from all SME financing documents, then automatically populating case files within their operational platform.

How the SME Lending Origination Accelerator works

  1. Request intake

    Funding requests are retrieved in real time from the customer service portal.

  2. Document processing

    Azure Document Intelligence extracts text and data from all attachments.

  3. AI agent

    A GPT-based agent interprets essential information from raw, unstructured data.

  4. Standardization

    Information is structured and standardized to ensure consistency.

  5. Case file

    Case files are created automatically in the company’s ERP.

By integrating directly with the company’s customer service portal (Zendesk), where loan applications were received and processed, the system retrieved funding requests in real-time. Through the development of a GPT-based agent designed to ingest, extract and interpret essential information from raw, unstructured data, the information was captured, it was standardized and structured, to ensure consistency and it was introduced in the company’s proprietary ERP.

Through the AI-powered platform developed in collaboration with Zitec, the lender managed to create a fully synchronized, end-to-end processing engine to automatically create comprehensive case files for SME loan applications.

Building on this foundational efficiency, the company is already moving toward a more sophisticated, self-reliant ecosystem in the upcoming project phases, shifting focus towards cost efficiency. The next stage of our roadmap focuses on financial intelligence. Through advanced Generative AI and Azure Document Intelligence, they plan to perform complex document analyses, for real-time insights into applicants backgrounds, in order to eliminate their reliance on costly external providers. This transition from third-party dependency to an in-house, AI-driven decisioning engine will solidify the company’s position as a truly autonomous, market-leading financial institution.

“Real impact when implementing AI technologies is about engineering a system that can scale and help an organization improve. By prioritizing a clean data foundation, they didn't just automate a workflow, they built an AI-ready organization prepared for the Agentic Era. This transformation positions the organization beyond the step of simple efficiency gains and into a position of long-term market resilience.”

Radu Săndulescu

Data & AI Director, Zitec

The Results

Through the AI-powered platform developed in collaboration with Zitec, the company achieved a 65%+ reduction in time spent on processing new funding requests, resulting in the ability to process almost 3 times more loan applications with the same team and with an almost 85% lower cost per application processing!

This high level of automation has boosted operational efficiency, allowing the team to focus on high-priority tasks while maintaining a smooth workflow. In day to day numbers, it means the automated analysis of over 350.000 pages of funding requests attached documents, and processing over 50.000 loan applications per month, while maintaining drastically accelerated loan eligibility response times, with negative replies delivered within minutes and positive replies provided within just 2 hours.

Furthermore, through the power of digital transformation, the company is constantly able to achieve a great risk-adjusted net interest margin, higher than the banking system average. This success is attributed to a strategic redistribution of working capital towards investments that offer a superior risk-adjusted return profile, resulting in higher profitability and greater financial stability for the business.

Before and after: impact of the AI-driven pipeline
Metric Before After Impact
Application intake Manual data entry and document review Automated extraction and case file creation 65%+ less processing time
Cost per application Manual processing and third-party OCR tools AI-driven in-house processing pipeline Almost 85% lower
Processing capacity Limited by team size Same team, automated workflow Almost 3x more applications
Eligibility response Slow, manual replies Negative replies within minutes, positive within 2 hours Faster answers for every SME