Your Last Administrator Quit at 50 Percent. That Wasn't a Failure. It Was the Design.

You know the moment. The settlement is signed. The administrator sends the notice packet. The first wave of claimants responds, and the completion number ticks upward in a way that feels like momentum. Forty percent. Forty-five. Forty-eight. Fifty.

Then it stops.

Not gradually. It stops like a car hitting a wall. The admin sends a reminder. Maybe two. The number barely moves. And somewhere around that point, the conversation shifts. Your administrator starts talking about "claimant engagement challenges." They suggest the firm do its own outreach. They forward you a spreadsheet of non-responders and, in the politest possible way, make it your problem.

You paid full freight for the half that was easy. Now you are staring at a 75, 80, or 85 percent release threshold, and the vendor you hired to get you there has effectively left the building.

That moment is not a fluke. It is what the architecture was built to produce.

The class-action hangover

Legacy claims administrators (Simpluris, Angeion, Epiq, the usual list) were built for class actions. In a class action, the job is: send notice, process claims, cut checks. "Send it and wait" works because opt-out rates are low and participation thresholds are rarely the bottleneck.

Mass arbitration is a different animal. You need individual claimants to sign releases, return documents, verify identity, and close. Every one of them. The administrator's job is not to notify a class. It is to convert a population, one person at a time, all the way to a signature.

Legacy systems were never designed for that. They have a notice engine and a portal. They do not have a completion funnel. So they do the part they were built for (contact and first response), and they stall right around the 50 percent mark because there is no second act in their playbook. In the matters we have studied and run, that cliff shows up with striking consistency.

Completion is a funnel, not a mailing

GroupSettle was built from the ground up for the post-50 percent problem. We treat every matter like a conversion campaign with multiple stages, not a single blast.

Here is what that looks like in practice:

• Native document signing. Release and closing statement happen in one session, with a full audit trail, no redirect to DocuSign, no extra vendor. The claimant gets a text, opens it, signs, and is done. Friction is where completion goes to die, so we removed it at the source.
• SMS and email sequencing. Not one reminder. A structured campaign that adjusts cadence and channel based on where the claimant dropped off. Built on Send It By Text's own infrastructure, not a licensed API with a per-message markup.
• AI super agent. Claimants have questions. "Is this real?" "What am I signing?" "How do I get paid?" In a legacy setup, those questions sit in a call center queue or go unanswered. Our AI super agent handles claimant inquiries over email and text, resolving more than 80 percent of them without human intervention, around the clock. That alone keeps the funnel moving when a traditional admin's office is dark.
• AI voice for holdouts. The last 10 to 15 percent of claimants are the hardest to reach. They did not respond to the text. They did not open the email. For those holdouts, we deploy AI-powered voice calls that feel human, not robotic, and walk the claimant through what they need to do. It is the layer legacy providers simply do not have.

The result, in the matters we run, is roughly 50 percent more effectiveness on completion compared to the legacy baseline. That is not a marketing number. It is what we observe when we compare outcomes on similar populations.

Less than half the price, and we only collect if you win

Here is where the story gets hard for a legacy provider to answer.

GroupSettle charges a flat $11.99 per signed claimant. The legacy rate runs $20 to $25. That is not a promotional discount. It is structural. GroupSettle is a division of Send It By Text, which means we own the SMS infrastructure, the e-signature engine, the voice platform, the KYC layer, and the disbursement rails. We do not license five tools from five vendors and mark each one up. We built the stack, so we run a lot of it at cost.

And then we go one step further: we work on contingency. GroupSettle invoices nothing until the firm's release threshold is met. If we do not get you to the number, you owe zero.

Think about what that means for incentive alignment. A legacy administrator gets paid whether you hit 50 percent or 85 percent. Their invoice looks the same either way. Our invoice only exists if you cross the finish line. Every dollar we spend on that AI voice call, every text in the drip sequence, every hour the super agent spends answering "is this a scam?" is a bet we are making with our own money that we can carry you to the number.

That is not charity. It is a bet we can afford because owning the technology makes the marginal cost of persistence low. We run the engine at or near cost, and our margin comes from winning, not from billing for effort.

The question is not whether your last admin tried hard enough

They probably did, within the limits of what they were built to do. The problem was never effort. It was architecture. A notice-and-wait system will always hit the same wall, because it was designed for a world where getting to 50 percent was the whole job.

Mass arbitration is a different world. The job is getting to 75, 80, or 85 percent. And that requires a fundamentally different machine.

We built that machine. We priced it at less than half the legacy rate. And we tied our fee to the only outcome that matters to you: hitting the threshold in your settlement agreement.