Dan Reid is not surprised that Xceptor, the data automation giant he formed two decades ago, finds itself at the vanguard of a change in the way financial institutions regard and use documents.
The rapid and accurate parsing of information from paper- and PDF-based reports has been made possible thanks to recent developments in artificial intelligence. The volume of documents that organisations need to process to meet modern business objectives plays to the strengths of Xceptor, which continues to invest heavily in AI applications to augment its automation capabilities.The UK-based platform provider was founded in the early 2000s to offer enrich, transform and load (ETL) capabilities built for financial institutions. It is primarily aimed at enabling middle and back offices of its 125 clients to design and manage data workflows without relying completely on their internal technology departments.
In the AI era, that has meant creating a bridge between their tech stacks and functionality of the hyperscalers like AWS and Azure. It removes the traditional choices that face institutions of building technical toolsets with multi-year implementation cycles or manual, spreadsheet-based processes.
“We were early to the table with AI capabilities and suddenly, many of the things that we do in Xceptor, like pulling data in from disparate, unstructured, untrusted sources, suddenly became something that we could enhance with tools provided by hyperscalers,” Reid told Data Management Insight.
“Therefore, the Azure Document Intelligence and Azure OpenAI capabilities have become a central part of what we do.”
Easy Fit
Xceptor hasn’t so much embraced AI as seamlessly slotted it into a business model that, in retrospect, appears purpose built to absorb the technology. For a company that had been created to provide the optimising connective tissue between organisation and emerging technologies, AI has been like an accelerant in Xceptor’s platform.Its latest AI developments comprise the use of agents that assist in data extraction on the Xceptor Data Automation Platform. That enhances one of the company’s key capabilities.
Reid says that AI has removed the distinction between document and data. Before AI, parsing information from unstructured formats was slow and often required manual input. Despite the troves of useable data within documents, the difficulty of fully extracting that information meant the processes that did so were of limited utility.
That has changed.
“There’s been a real shift to this kind of recognition that documents are data –?a document isn’t a different thing from data; it IS data,” Reid said.
“And so there’s been this big harmonisation between data in its conventional sense, data in a database, data in an application and documents coming together to be seen as two parts of the whole, where that whole is the business intelligence that’s required to run your operation effectively.”
Regulatory Change
A leading driver behind Xceptor’s growth has been regulators’ requirements that financial institutions streamline their reporting, settlement and other processes to ensure stable markets and to improve market efficiency. Ever-faster trade settlement times have been demanded even when many organisations have been unable to meet those demands on their own.
Reid expects the coming years to see more of the same, especially with T+1 expectations looming.
“The shift to T+1 has profound implications on all of the processes that exist upstream of that settlement,” he said. “The days of organisations being able to book a trade and then have some kind of fragmented, ad hoc, inefficient, poorly automated processes between there and settlement, are gone because you can’t reliably settle on T+1 if you’ve got poor processes” in your tech stacks.
Focused Development
The Xceptor platform was built for financial services. The company could have taken multiple paths in developing its business plan – healthcare and manufacturing were candidates – but Reid decided to concentrate on financial markets because that was where he saw greatest need.
“We made the conscious decision that the intensity of data challenges that exist in capital markets is sufficient that we want to focus there,” he said.
“So we tackle the hardest data automation challenges, and we do so end-to-end. That means embedding AI in parts of that process, but it also means recognising that AI can’t solve every problem. And even if it can solve some of the problems, it might not be the best solution to those problems.”
Despite the automation strides that Xceptor has brought about in the past two decades, there still remain lots of “low-hanging fruit” for it to address. Some organisations are still using manual processes in their day-to-day operations where technology can take the burden, especially in data extraction.
For those – and all of Xceptor’s clients generally – AI has hastened the need for change, argued Reid.
“The pace at which AI is moving in the market means that our customers need to be in a position where they can bring in new capabilities really quickly,” he said. “And the rationale for the pieces that we’re adding is to enable our customers to do that as effectively as possible. So we’re supporting customers in moving as fast as they want and as fast as they need, to keep pace with the market.”
Subscribe to our newsletter
