Lexitas’ drive to modernise its finance operations has turned one of its most labour-intensive processes into an early showcase for agentic automation.
The company, which provides technology-enabled litigation support services and has grown through more than 53 acquisitions since its founding in 1987, had accumulated the kind of operational sprawl that often follows rapid deal-making. Boomi said Lexitas has used its platform to unify data from more than 40 acquisitions, reduc...
Continue Reading This Article
Enjoy this article as well as all of our content, including reports, news, tips and more.
By registering or signing into your SRM Today account, you agree to SRM Today's Terms of Use and consent to the processing of your personal information as described in our Privacy Policy.
e duplicate records by 80% and shorten its monthly close by 36%. That broader data work created the foundation for a new AI agent designed to tackle lockbox payment processing, an area that had become too complex for manual handling alone.
Each day, more than 2,500 payment lines arrived through Lexitas’ main lockbox, often with customers paying several invoices in a single lump sum and providing little or no explanation. Matching those payments required staff to trawl through emails, enterprise systems and legacy records to determine which invoices had been settled. According to Lexitas, the task had once consumed a team of 15 to 20 specialist accountants.
The push came from finance rather than IT. Chief Accounting Officer Sherry Bourque identified the process as ripe for automation and worked with John Baker, Lexitas’ chief information officer and chief information security officer, to turn the idea into a live system. Baker has said the company had already been using Boomi for around five years to manage integrations across acquired businesses, which made it a natural partner for the project.
Boomi ran a proof-of-concept workshop at no cost, with the pilot beginning in mid-November 2025 and moving into full production by February 2026. Baker later said the team had initially approached AI cautiously, given the number of warnings surrounding the technology, but the staged rollout helped build confidence.
The system uses a multi-agent design orchestrated through Boomi AgentStudio. Separate agents review the shared inbox, parse lockbox files and query the ERP and other internal systems, then compare the results to decide how a payment should be applied. Anthropic’s Claude, accessed through AWS Bedrock, provides the language model layer, while Amazon Textract handles document recognition and Boomi Data Hub supplies the governed data foundation.
Payments that are matched with confidence flow straight into the ERP. Cases that cannot be resolved with certainty are routed to the accounts receivable team, along with a summary of the agent’s reasoning. Lexitas says the process keeps accounting in control throughout, with a full audit trail and oversight tools in place.
Governance was central to the design. Baker, who holds responsibility for both technology and security, required that payment data stay within the Lexitas environment and not be used for external model training. Boomi Agent Control Tower provides monitoring, alerts and an immediate shut-off mechanism if behaviour strays from agreed limits.
The early results have been substantial. Lexitas says the agent now handles 46% of daily payments, with 30% processed fully without human intervention and a further 16% completed with human assistance. The company also says the system has improved auto-posting rates, accuracy and exception handling, while reducing the repetitive workload previously shouldered by finance specialists.
The initiative is already expanding. Lexitas is now extending the framework to credit card processing, a second bank’s lockbox, accounts payable and vendor payments. Baker has argued that the best starting point for AI is a narrow, measurable business problem, rather than a broad and abstract ambition, and says the biggest challenge is usually not building the agent but securing the right data and controls.
Boomi is meanwhile using the Lexitas case to illustrate a wider theme in enterprise AI: that useful automation often depends less on model sophistication than on data quality, integration and governance. For organisations with fragmented systems and multiple inherited processes, Lexitas’ experience suggests that the path to AI value may begin with finance, but it quickly becomes a broader exercise in operational discipline.
Source: Noah Wire Services
