The following article was written in association with Applied Systems.
Matt Lane (pictured) arrived in the insurance industry three years ago from travel technology. His first observation was not about the complexity of the products or the structure of the market. It was about the people.
"A fantastic industry with people that really want to work together," he said in an interview at Applied Net 2026 in Washington DC last week. "But still pretty early days in many ways in digital transformation, if I'm being honest."
That assessment shapes everything Lane is building as senior vice president of product at Applied Systems, where he oversees Epic, the company's flagship agency management system. His starting point is not what the technology can do. It is what brokers are actually doing every day - and how much of it should not require a human being at all.
The numbers Lane works with give his product decisions a particular weight.
Epic has 200,000 active users. Each spends roughly two hours a day actively in the system. Across that base, users conduct more than 200 million searches per year - locating policy records, client information, documents buried in attachments, scanned files that previously required opening individually and searching manually.
Lane uses Pendo, a product analytics platform, to understand precisely where those hours are going. The data consistently shows three concentrations: brokers spend the most time reviewing and validating policy information, creating activities and tasks to document their work, and searching within the system.
All three are areas where technology can reduce the burden significantly. All three are areas where Applied has active development underway.
The new search capability - a Google-style bar present on every screen in Epic, indexing records, attachments, and scanned documents - is currently in pilot with approximately 40 customers. The user response has been striking. One told Lane it was the best feature they had seen in their career. Another said they could not have reached the following year without significant additional hiring if the tool had not been available.
"Is search some breakthrough product strategy idea?" Lane said. "Not really. Is delivering it really difficult? Yeah, it's really hard."
The difficulty Lane is referring to is not conceptual. It is the engineering required to make 200 million annual searches across a base of 200,000 users feel instantaneous, to surface content from scanned documents as reliably as from structured records, and to do it without disrupting a system that brokers depend on for their most critical daily work.
Search is the most visible example of where the broker experience is being rebuilt. The submission workflow is where the industry's underlying structural problem sits.
Epic's 900 screens of forms have historically been filled out manually. ACORD forms in the US, CCO forms in Canada - documents that can run to hundreds of fields. Lane's annual voice-of-market survey has returned the same response for years: help me enter data without rekeying it, check the accuracy of what is in the system, make it easier to find things.
"The product strategy part of my job is not that complex," Lane said. "Building products that can do this at scale, getting people to change their behaviors - even when you offer them something great - is really, really difficult."
Autofill addresses the data entry side directly. The capability extracts information from source documents and populates submission fields automatically. In the employee benefits space, where Autofill is already live, tasks that previously took 30 minutes are now handled without manual input. In some cases, Lane noted, data that was not being captured in the system at all before - because the manual effort required to record it acted as a barrier - is now populated automatically.
The more ambitious step is connecting Autofill to the carrier-side intelligence that Cytora provides. Rather than requiring brokers to review every field in a populated submission for accuracy, the combined approach surfaces only the data elements a specific carrier actually needs to issue a quote.
"Maybe we don't need people to review 100 data elements," Lane said. "We just need them to review the 10 that are actually relevant."
Cytora's technology is embedded in Epic under the Autofill and policy checking product names rather than its own brand. Lane is clear about why. "My customers buy Epic," he said. "They care about the technology being absolutely rock solid."
The EZLynx personal lines rating engine integration is the clearest illustration of the gap between owning a technology and delivering it to users.
Applied acquired EZLynx five years ago. Its personal lines rating engine is, in Lane's description, the best in the United States. Epic has had its own personal lines rating capability throughout that period. The decision to replace the internal capability with the superior acquired one - embedding EZLynx's engine directly into Epic's workflows - seemed straightforward in principle.
"This should be easy because we own these things," Lane said. "Boy, was I wrong."
The work required deep workflow embedding, user provisioning and deprovisioning systems, and architecture alignment between two products that had developed independently for five years. The integration is now in pilot with seven customers and has processed approximately 1,400 quotes, with strong early reception.
The lesson Lane draws from it is broader than the EZLynx case. Acquired companies bring technology, people, and external perspectives that a large incumbent cannot easily grow organically. The entrepreneurs who built those companies often become peers who challenge assumptions from inside. The value of the acquisition strategy is not just the product - it is the caliber of the people who built it.
"I care a lot about having entrepreneurial, cutting-edge technologists that can sit side by side with me who have outside perspectives that Applied might not be able to grow organically," Lane said.
Post-bind data flows in the insurance industry - the Ivans rails that carry policy details, commission information, and documents between carriers and brokers - work reasonably well by Lane's assessment. The harder problem is what happens upstream, before a policy is bound, particularly in commercial lines.
Many brokers still prepare a commercial submission in Epic, print an ACORD form to PDF, and email it to carriers. The digital connection that would allow that submission to flow directly from the management system to the carrier underwriting workflow does not yet exist at the scale the industry needs.
Submission Manager takes an early step toward it. The product provides visibility into both Epic-originated submissions and emailed ones - giving brokers and carriers a shared view of where submissions stand, regardless of how they were sent.
"You have what I view as like the beginning feeling of a digital marketplace on the commercial side," Lane said. "But I don't think we've done everything we need to do to truly connect the data and have a shared record of risk. It's still disparate, to be honest. So there's work left to do."
The most significant idea Lane raises is not a product that exists today. It is the destination he is building toward - and the framing he uses for it is the most useful way of understanding why the submission, search, and connectivity investments matter as a connected programme rather than separate features.
Insurance is a record-keeping business. The agreements between carriers and insureds are legal contracts, and the broker's job is to advise and protect clients on the basis of what those contracts say. The problem is that the same information exists in multiple places simultaneously - a policy document, a carrier download, a publicly available data source - and those versions do not automatically agree with each other.
A single carrier may send commission statements in up to 12 different formats. Data that was accurately recorded at binding may drift from the carrier's own records over the life of a policy. The broker working from their management system cannot always know whether what they are looking at reflects current reality without manual verification.
Brokers are currently spending millions of dollars with business process outsourcing firms to do that verification work. Lane's view is that it should not require human labor at all.
"I want to figure out how we give our customers incredible confidence that the risk profile, the customer information, and the coverage information they're looking at is constantly self-healing and accurate," he said. "Maybe the document says this, the carrier download says that, information on the publicly available internet says something else. Where are the actual truths in this information?"
The five-year vision is a data layer that reconciles those sources continuously - not a static record that is accurate when entered but that drifts over time, but a system that actively maintains the accuracy of what it holds by cross-referencing everything it knows.
Autofill and policy checking are the first expressions of that direction. They reduce the manual burden of getting data into the system accurately in the first place. What Lane is building toward is the next step: a management system that does not just record information when brokers enter it, but that continuously validates whether what it holds is true.
For a broker whose clients rely on the accuracy of that information - sometimes at the moment of a claim, when the gap between what the system says and what the policy actually provides matters most - that is not an incremental product improvement. It is a different kind of infrastructure entirely.
