SlateWorks wins ABBYY hackathon with mortgage AI tool

ABBYY has named SlateWorks Global Solutions the winner of its developer conference hackathon in Bangalore. The winning entry focused on mortgage document review.

SlateWorks built a tool called ABBYY Intelligent Mortgage Auditor, or AIM, which uses ABBYY's document processing software alongside generative AI and retrieval-augmented generation to automate checks on mortgage files. The system reviews loan applications, tax transcripts, automated underwriting system reports, bank statements and related records to identify mismatches, anomalies and compliance issues.

The project was selected from more than 25 teams in the competition. The other finalists were Tata Consultancy Services, Alltius, Proservartner, Slateworks Global Software, Tangentia, Ellucian and Tech Mahindra.

Judges cited the project's readiness for deployment and its focus on a routine but heavily regulated lending task, where document verification is often still handled manually. The software can also recommend operational actions and includes a conversational interface for mortgage auditors working to government-sponsored enterprise and investor guidelines.

The result highlights a growing commercial use case for document-focused AI in financial services, where lenders are under pressure to cut processing times without weakening audit trails or compliance checks. Mortgage reviews typically involve large volumes of paperwork from multiple sources, making them a target for automation vendors.

SlateWorks developed the winning system with input from internal subject matter experts. Some members of the team had also taken first place in the previous edition of the hackathon, according to ABBYY.

"At SlateWorks Global Solutions, we consistently see the significant effort organizations spend reviewing documents, verifying information, and maintaining regulatory compliance, work that is critical yet still largely manual," said Sam Kharidi, Founder and CEO, SlateWorks Global Solutions.

"Winning the ABBYY Ascend DevCon Hackathon validates our vision that combining deep domain expertise with ABBYY's document intelligence and AI can create solutions that extend far beyond traditional document automation. We envision the platform powering intelligent solutions across lending, banking, insurance, healthcare, supply chain, finance, government, and many other industries where documents drive critical business decisions," Kharidi said.

The winning team received cash prizes, automatic entry into ABBYY's Most Valued Partner program and a place in its Innovation Hub, which showcases new projects and experiments from partners and developers.

Hybrid approach

The hackathon was part of ABBYY's annual Ascend developer conference, which drew record attendance from partners, developers and corporate users across sectors including finance, manufacturing, healthcare and retail, according to the company. ABBYY said attendance has quadrupled since the event began four years ago.

A central theme at the gathering was a hybrid AI model that combines deterministic document processing with generative AI. ABBYY said the approach is intended to help businesses decide when to rely on structured extraction techniques and when to apply generative reasoning, while keeping processes auditable.

That message reflects a broader shift in the enterprise software market. Many companies are moving from pilot schemes to operational deployments and are under pressure to show that AI tools can be governed, monitored and integrated into routine workflows.

Conference sessions focused on document AI, agentic workflows and data quality. The aim was to show how organizations can connect different AI methods in a single pipeline while preserving controls around validation and orchestration.

Neil Murphy, Chief Revenue Officer at ABBYY, said businesses are no longer satisfied with experimental systems. He linked that shift to demand for AI tools that can be measured against production standards.

"The conversation has shifted, and organizations are moving beyond AI experimentation and demanding solutions that are accurate, auditable, and ready for production," Murphy said.

"The innovations showcased at this year's DevCon demonstrate how a hybrid approach of purpose-built document AI and generative AI can work together to solve complex business problems at enterprise scale," he said.

ABBYY also used the event to demonstrate its FineReader Engine and optical character recognition technology, which it said processed documents into structured DocLang output at 500 pages per second. The demonstration was presented as evidence that document-heavy AI systems can be designed to handle high-volume workloads while meeting accuracy and compliance requirements.

ABBYY, based in Austin, Texas, says more than 10,000 enterprises use its software. The company focuses on document AI and automation tools that convert business records into structured data for downstream processes.