Lucite CEO sees a bright future for AI in benefits claims but not without its limits
Before co-founding Lucite, an AI platform built for Canadian benefits advisors, Jason Spector was a technology investment banker at J.P. Morgan in New York and then a growth equity investor backing cybersecurity and fintech startups.
But the spark for Lucite came from the tedium of that work, where much of his time was spent moving data from PDFs into spreadsheets and slide decks. He recruited a college friend, Richard Greenbaum, a former machine learning lead at Meta, and pitched him on the problem.
They went through Y Combinator, quit their jobs, and spent six months selling into finance. The pivot came when an investment banker introduced them to a Canadian group benefits advisor who had exited his firm in 2018.
“He told us, ‘You have no idea how much this is needed in my industry,’” recalled Spector, who's also Lucite's CEO. “He explained the pain points and how benefits advisors are doing so much manual work to move unstructured data from insurance carrier documents into reporting and analyzing that information. All of our light bulbs went off of when we realized this is the perfect use case for what we built.”
Spector is quick to highlight how Lucite is already delivering value across the main tasks benefits advisors handle, pointing to four areas in particular: plan design, marketing and quoting, renewal reporting, and claims experience reporting.
He believes these are exactly the kinds of workflows where automation can make a practical difference, because they involve large amounts of manual review of carrier documents, contracts, and plan details.
Where advisors previously had to gather documents from multiple carriers, sift through them by hand, pull out the relevant details, and then turn that information into recommendations and client reports, Lucite is designed to replace much of that process with a simpler upload-and-extract workflow, allowing advisors to move much faster from raw documents to usable analysis, explained Spector.
He argues that the biggest gain for Lucite comes in claims experience analysis.
“I see the biggest impact in claims experience and analysis and looking at all of this data and information because you're able to structure it and store it in a secure way. Previously, all of this just lived in scattered files, across many different files that looked like different things, and it was very hard to do any sort of analysis on these documents,” Spector explained.
Spector describes Lucite as being organized around client groups, with the platform built to mirror the way benefits advisors already work. When a client is nearing renewal, an advisor typically gathers the current carrier’s renewal report along with a stack of claims documents from the past year.
Those files are then uploaded into the client account in Lucite, where the system processes them and turns the information into dashboards, charts, draft reports, and AI-generated insights that highlight trends, year-over-year changes, and issues worth raising with the client.
Spector noted the aim is to take information that was previously locked across multiple PDFs and convert it into a branded report that is largely ready for client delivery. The same logic applies when advisors shop the market. Instead of manually comparing carrier proposals one by one, the platform helps generate proposal requests, ingests the responses, and then produces side-by-side comparisons of pricing and plan differences.
Lucite isn’t trying to be a general-purpose AI tool. Instead, it’s built as specialized software for a narrow market, shaped around the day-to-day workflows of benefits advisors, said Spector.
He argues that the difference lies not just in pulling information from documents, but in turning that information into something advisors can actually use. The platform stores data securely, shows historical information across client groups, and feeds it directly into branded, white-labelled reports tailored to each firm.
That makes Lucite more practical than general AI tools, which may summarize documents but do not solve the full workflow, said Spector.
He also points to concerns that general tools cannot fully address, especially around security, data residency, accuracy, and traceability, particularly as advisors want to know where their data is stored, whether it remains in Canada, and whether they can quickly trace extracted information back to the original document.
Spector noted Lucite was designed to answer those questions while also fitting into firms’ existing reporting templates. He also underscored security sits at the center of Lucite, pointing to measures like SOC 2 Type 2 compliance, ongoing auditing and penetration testing and PIPEDA compliance, notably as benefits advisors are dealing with personal and health information.
He also stressed that the product is shaped by close contact with customers. Lucite’s engineers are kept in-house rather than outsourced, and they regularly speak directly with advisors to hear feedback and identify pain points. That hands-on approach is as important as the software itself, because it lets the company keep refining the product around the real problems advisors are trying to solve, he underscored.
Still, he’s also aware of where Lucite falls short. The most common request the company hears is for automated document retrieval from carrier portals. And with more than 10 carriers in Canada, each with different systems and varying levels of complexity, advisors often have to navigate multiple steps just to find and download the right claims reports for a single client. Furthermore, those documents may be spread across health, vision, and dental.
"Today, Lucite doesn't go and automatically pull those documents, that's on the advisor to go get those documents," he said, noting it’s a problem the company is aware of and in discussions with carriers about, but remains unsolved for now.
Looking ahead, Spector sees Lucite moving beyond document extraction and reporting into analysis and recommendations. With years of historical claims and renewal data stored per client, the platform could begin forecasting outcomes.
"We could actually start predicting what a fair renewal can be, and I think those predictive analytics is where I see a lot of the next phase of our platform," he said.
While Spector does see AI reshaping white-collar work broadly, he resists the more extreme predictions, emphasizing that it won’t eliminate every role.
"I don't believe it's going to wipe out every single job, but I do believe it's going to significantly impact organizations and the way they think about doing this type of work," he said.
That shift is already underway. For benefits advisors, the hours spent on manual data entry and reformatting are the first to go.
But efficiency alone is not the point. What advisors care about most, Spector said, is growth and better client service, highlighting stronger analysis, sharper reporting, and more informed advice on plan management and cost control. After all, benefits sit as the second-largest expense behind payroll for most businesses, which makes the advisor's role critical, noted Spector.
"All of those hours that are spent copying and pasting data from one document to another file and formatting it are going to go away," he added.
