Our Story
When Jason (our CEO) began his career as a corporate lawyer in 2007, one of his first assignments was performing a definitions check on a loan agreement, which was to find "Defined Terms" that were never used and to find Capitalized terms that were never defined. He understood the necessity of the task, but was surprised that no standard procedure existed to perform the task.
One evening, around 9PM, a request came in to perform the definitions check, and to have the check completed and reflected in the upcoming draft in the morning. Due to other pressing assignments, Jason started the check around 2AM, and finished it around 4AM. He had submitted the work, but it was too late; the revised document that came out in the morning did not reflect Jason's findings. After being yelled at, he wondered if someday this process could be automated, similar to the way redlining or spell checking had become automated.
When Jason decided to leave law practice to start Celant Innovations, he revisited pain points of modern day corporate practice. After learning that the latest technology had not flowed through the existing solutions for defined terms checks, the idea of DefCheck.com was born. After continuous refinement over 18 months, the algorithm produced reliable results on actual commercial agreements and the patent application was filed, and was granted the patent (no. 10713482) from the USPTO on July 14, 2020.
DefCheck.com is a simple and reliable tool to automate the defined-terms related error detection process. Alongside the fully highlighted document — Green (everything OK), Red (defined term not used), and Pink (Capitalized term not defined) — DefCheck now produces a prioritized findings report, so confirming the errors takes a practitioner minutes, not hours.
What Makes DefCheck Different
It is patented. The core detection method was granted U.S. Patent No. 10,713,482 in 2020 — the product of 18 months of refinement against actual commercial agreements before launch, and continuous improvement since.
It is measured against real lawyer work, not synthetic tests. Every change to DefCheck's engine is benchmarked against a corpus of SEC-filed credit agreements, merger agreements and indentures that were reviewed line-by-line by a practicing securities attorney. On that benchmark, DefCheck surfaces over 98% of the issues the human review found — while filtering out the noise (tables of contents, section headings, signature pages) that makes naive term-checkers exhausting to use.
It has read more contracts than any associate ever will. DefCheck's suggestions draw on a research dataset built from 7,200 SEC-filed contracts, in which human reviewers individually examined more than 30,000 flagged terms and recorded the correct resolution for each. That is why, when your draft says "Issuing Bank" but defines "Issuing Lender," DefCheck doesn't just flag the term — it tells you which defined term you probably meant, and how often that exact slip has appeared in other agreements.
It is built for the AI era. Every check produces an analysis report designed to be pasted directly into an AI assistant such as Claude or ChatGPT, with the context, evidence and instructions the AI needs to act like a trained reviewer. In our blind tests, an AI assistant working from DefCheck's report reproduced roughly nine out of ten items on the expert attorney's final review memo — including some the expert missed.
About Celant Innovations
Celant Innovations designs, develops and operates cloud-based and on-premises solutions relating to legal document drafting and analysis. We strive to improve efficiencies in traditional processes, focusing our expertise in legal, accounting and financial industries. Our goal is to create easy-to-use products that resonate with our customers.