The Last 15 Percent of Your Claimants Won't Read a Text. They Won't Open an Email. Now What.

You know the moment. Your completion rate climbed fast in the first two weeks. Texts went out, emails landed, claimants signed. The dashboard looked promising. Then it stalled. Not at the finish line. Somewhere around 60 or 65 percent. Your administrator sent another round of emails. Maybe a reminder text. The number barely moved. And you sat there looking at a threshold of 80 percent, doing math you didn't want to do, wondering whose job it was to close that gap.

I've heard this story from enough partners to know it isn't an outlier. It's the pattern. And the reason it keeps happening is that legacy administrators don't have a channel for the people who won't engage with text or email. They have two tools. When both fail, they shrug. That's where your matter sits.

The Holdout Isn't Ignoring You. They're Ignoring the Channel.

There's a category of claimant that every mass arbitration matter produces. They signed up. They're entitled to their share. They're not hostile. They're just not going to tap a link in a text message or click through a five-paragraph email. Maybe they don't trust the sender. Maybe they skimmed it and meant to come back. Maybe they're 68 years old and their inbox is a graveyard. Whatever the reason, the channel failed, not the intent.

Legacy administrators treat this group as a loss. A shrinkage line. The claimants who "didn't complete." But in a matter where every percentage point between 65 and 80 determines whether you hit your threshold, those aren't acceptable losses. They're the whole ballgame.

AI Voice Exists Because We Needed a Third Channel

When we built the completion engine at GroupSettle (through Send It By Text, which owns every layer of the stack), we started with native document signing. Release, closing statement, identity verification, all in one session. No third-party e-sign vendor bolted on. No extra click that loses the claimant between tabs.

Then SMS and email, because those are the fastest channels to reach people at scale. Then the AI super agent that handles 80-plus percent of claimant inquiries over text and email, answering the questions that would otherwise stall someone mid-process.

But we kept seeing the same ceiling. A reliable cohort of claimants that none of those channels could move. So we built AI voice into the stack. Not a robocall. Not a pre-recorded message telling someone to visit a website. An AI-driven outbound call that can confirm identity, walk through the release, answer questions in real time, and get a claimant to completion on the phone. Same audit trail. Same compliance framework. Just a different door into the same room.

We own the voice layer the same way we own SMS, email, signing, and the super agent. It's not a vendor integration. There's no per-minute surcharge from a third party. It's just another channel in the funnel, and it exists specifically for the holdouts that every other administrator writes off.

This Is Why Legacy Admins Stall at 50 Percent

The 50-percent cliff isn't a mystery once you understand the toolset. Legacy claims administrators were designed for class actions, where the job is "send the notice and wait." In mass arbitration, that approach gets you to roughly 50 percent completion. The claimants who are easy to reach and motivated to act will complete on their own. Everyone else needs persistence, multiple channels, and someone (or something) to answer their questions at 10 p.m. on a Tuesday.

Legacy providers don't have that persistence because they don't own the tools. They license SMS from one vendor, e-signature from another, maybe a call center from a third. Each integration is fragile, each vendor adds cost, and none of them were built to work together as a completion funnel. So the administrator does what it can with what it has, invoices the firm for the full campaign, and moves on. The firm is left to drag the matter from 50 to 75 or 80 or 85 on its own.

GroupSettle doesn't work that way. In the matters we run, we see roughly 50 percent more effectiveness on completion compared to the legacy model. AI voice is a big part of why. It's the layer that turns a stalled campaign into a moving one, because it reaches the people that text and email simply cannot.

Less Than Half the Price. And We Only Get Paid If You Win.

Here's the part that tends to make partners pause. GroupSettle charges $11.99 per signed claimant. The legacy rate is $20 to $25. That's not a promotional offer. It's the structural result of owning the entire stack instead of licensing five vendors and marking each one up.

And the fee is contingent. GroupSettle invoices nothing until the firm's release threshold is hit. Miss the number and the firm owes zero. That's not a slogan. It's the pricing model. We can afford to make that bet because we run the engine at or near cost (owning the technology makes the marginal cost of each additional channel, including voice, very low) and because we only take matters we're confident we can carry to completion.

When your incentive is identical to the firm's, you build tools like AI voice. You don't write off the last 15 percent. You go get them.

The Channel You're Missing Is the One They'll Answer

Every matter has holdouts. The question is whether your administrator has a plan for them or just a report showing they didn't complete. GroupSettle was built for the second half of the campaign, the part that actually determines whether you hit your number. AI voice is one layer of that. The owned stack, the native signing, the super agent, the contingency model: they all connect. None of them work in isolation, and none of them exist at a legacy provider.

If your current administrator's answer to holdout claimants is "we sent another email," you already know how that ends.