Your Next Great Hire Doesn't Need a Desk

There's a wave of new AI tools landing in insurance. Point solutions targeting gaps the AMS has left open for a decade: proposals, policy checking, commission statements. The good ones genuinely deliver. They're hyperfocused on one neglected problem, and they finally solve it.

But each one is its own silo. Its own login. Its own data. Its own usage-based subscription, metered by the transaction. Every gap you close adds another tool to learn, another bill to pay, another set of records that doesn't talk to anything else in your core systems.

The streaming-service problem has made its way into your agency. We're all living it right now. One service used to have everything; now every studio has its own, and you're paying for five subscriptions just to watch all your favorite shows. I've lost count of how many times I've cancelled one only to add it back a month or two later because a new show I actually want to watch drops on it. Looking at you, HBO Max. AI tools in insurance are sprinting down the same path. A new tool for every gap. A new bill for every tool. Data scattered across all of them, none of it connecting.

Stop thinking of AI as another subscription. Think of it as someone you hire.

Hire an AI agent like you would a human

When you bring on a new hire, you give them a role. You teach them your voice and your standards. You hand them your SOPs, give them access to your systems, put a supervisor over them, and give them time to ramp.

An AI agent works the same way. The outcome is the same: someone who shows up every day and does the job.

That changes three things at once:

Your customers meet someone. A named agent with a consistent voice doesn't read as faceless. It reads as a representative of your brand who happens to be available at 11 p.m. on a Sunday. No off days. No post-lunch slog.

Your team stops thinking "the system" and starts thinking "my colleague." No more "should I even use this? It charges us whether it works or not." It's a subtle shift, but it's the one that drives adoption. People route work to a teammate.

You can actually hold it accountable. A named role with defined responsibilities can be measured, coached, and improved, the same way you'd manage anyone's performance. "How did intake go this week?" becomes a real question with a real answer.

Start at the front door

The unlock here is massive. The only question left is where to start.

For most agencies, the answer is the front door. Your website.

It's where first impressions are made, and where opportunities quietly die. Your website's whole job is to turn a visitor into a conversation. Yet that's exactly the moment most agency sites drop the ball.

Look at the "Get a Quote" experience on most agency websites and you'll see the problem in its purest form. There's a "contact us" form, sure. But there are also Get a Quote forms for every line the agency writes: Personal Lines, Commercial Lines, Employee Benefits, and whatever specialty products round out the book. Each one is a clunky form tree the prospect has to click through, field after field, only to land on the same dead end every time: "Thanks, we'll contact you later." Be honest: nobody actually sends a customer there. It's a box you check so you can say you have it on your website. A digital obligation, not a destination. Your real business happens by phone, by email, by referral. The website is just there.

That's a missed opportunity hiding in plain sight. Stop checking the box. Your website is your digital front door, and every front door deserves someone standing at it, greeting people on their way in. That's where your agent lives.

This is exactly the seat a branded AI agent should fill first.

Instead of a cold form tree that ends in "we'll contact you later," your client meets your agent. On brand, on message, and working around the clock. The agent greets them, helps them start the application, asks for the supporting documents it needs, and fills out the form for them as they talk. Then it hands them off to the right person with a warm, high-momentum handoff: a hot lead that arrives already organized and ready to quote, not a cold name in a queue. The intake that used to take days of back-and-forth happens in a single conversation. Nothing gets dropped. Nothing gets re-keyed three times. And the very first interaction a prospect has with your agency reflects who you actually are, rather than the limitations of an off-the-shelf web form.

That's not a downgrade in service to save money. It's an upgrade in every direction at once: faster response, cleaner data, zero dropped opportunities, and a first touch that finally matches the quality of everything that comes after it.

The front door is the natural starting point because it's visible, it's measurable, and it pays for itself almost immediately. You don't have to reorganize your whole operation to feel the difference. You just have to let your new hire start answering the door.

And once you're comfortable there, the sky is the limit. The front door is where you start, not where you stop. The same agent that captures a quote request can go on to handle the quoting itself, build the proposal, take on servicing, and own every moment in between. You earned its trust at the door. Now give your agent a promotion.

Meet Harper

This is easier to picture than to describe, so let us introduce you to Harper.

Harper is the agent we built at Crux to show agencies exactly what this looks like. By now, your customers and your team are already using ChatGPT and Claude every week. They know what a real AI conversation feels like, and they know how capable these reasoning models have become. Harper brings that same caliber to your front door: a clear, professional voice that sounds like she belongs to your agency, not a generic chatbot script. She asks the right questions in the right order, captures clean and structured information the first time, knows what she can handle on her own and when to bring in a human, and hands everything off to the team without a single detail falling through.

The name, the voice, the defined job are what an agency sees and stands behind. Harper isn't "the system," she's a teammate you'd be comfortable putting in front of your best client. But the reason she actually works is the technology underneath, and that's not something we hide. We built an insurance-focused agent harness underneath: a supervisor that keeps her on task and in bounds, branding that makes her unmistakably yours, and the tool access to do real work inside the systems your agency runs on. That harness is the difference between a chatbot that talks and an agent that gets the job done.

Harper is our demo, but the point is what she represents: your agent won't be Harper. It'll be yours. Your name, your voice, your standards, doing the work the way your agency does it. Harper just shows you how quickly that becomes real.

It's not just for your customers, it's for your team

Here's the part most people miss when they evaluate AI agents purely as a customer-facing thing.

The biggest beneficiaries are usually on the inside.

Plenty of agencies still want the producer or account manager leading the relationship, and that's exactly right. In those cases Harper turns inward and works for your staff: the producer or account manager uploads the existing policy and supplemental documents, and Harper fills out the application from them and gets the whole thing organized and ready to submit for quote. Same agent, same standards, just pointed at your team instead of your customer.

There's another side to having Harper on the inside: connecting her to the systems you already run on. Once she's connected, Harper can answer the questions that today mean logging into your software, clicking around, and running a report just to find one piece of information. Instead of going to look it up yourself, you ask Harper. And she answers.

A well-defined agent does the setup work so your people don't have to. It's not there to replace the producer's relationship or the account manager's judgment. It's there to do the legwork that gets a submission ready, so your team can spend their time on the client and the deal. Your branded AI agent isn't only a better front door for customers; it becomes a teammate your own people lean on every day.

When the agent handles the front door well, or works alongside the person who does, the people behind it get to do their best work. That's the whole point.

A teammate that only gets paid when it does the job right

Here's where the "hire" framing stops being a metaphor and starts being the business model.

Think about how most AI is sold today. You pay for tokens. You pay per transaction. You pay for usage, and you pay it whether the work was done well or not at all. The meter runs the same whether the agent solved the problem or spun in circles, whether the submission went through clean or landed in someone's lap to redo. You're paying for activity. You're hoping for outcomes.

That's backwards. And it's not how you'd ever pay a person.

You don't pay a great hire to look busy. You pay them to get the job done. So that's how Crux works: your branded AI agent doesn't get paid unless it does the job right. Pricing is tied to outcomes (the submission completed, the certificate issued, the renewal handled), not to how many tokens it burned or how many times someone clicked.

This does something subtle and important: it puts everyone on the same side of the table. When we only get paid for outcomes, we're not incentivized to drive up your usage. We're incentivized to make your agent better. Faster, sharper, more accurate, handling more of the job with less friction every month. Your success is the only thing that pays. So we're all pulling in the same direction: yours.

A usage model rewards consumption. An outcome model rewards results. Only one of those is aligned with how you actually run your agency.

The future of work isn't coming, it's here

It's tempting to treat all of this as a someday problem. A "we'll get to it after Q4" problem.

It isn't. The shift is already underway, and the gap between agencies that have made it and agencies that haven't is widening every quarter. The ones who win won't be the ones who bought the most AI tools. They'll be the ones who learned how to work alongside AI: who hired it, named it, gave it a role, and built their operation around what that unlocks.

The way we work is changing. Not someday. Now.

Soon, your AI agent will be talking to theirs

Here's the part that should move this up every agency's priority list.

It's not just that you'll have an AI agent. It's that everyone you do business with will too, and those AI agents are starting to talk to each other directly. AI-agent-to-AI-agent commerce has moved from whiteboard to production: in late 2025, Visa introduced the Trusted Agent Protocol with more than a dozen partners, and Mastercard launched its own agent protocol the same quarter. These are open frameworks that let AI agents identify themselves, prove they're authorized, and complete payments on behalf of the business they represent, all governed by signed, auditable limits.Visa, Oct 2025American Banker, 2026 These AI agents are being credentialed and trusted to transact. That's not a 2030 forecast. It shipped.

Inside insurance, the same shift is well underway. Allstate has publicly launched its own AI agent (an "AI Sidekick" the CEO has discussed on the company's 2026 earnings calls) and plans to scale it across roughly 20,000 sales and call-center representatives.A.M. Best, 2026 Vertafore and Applied are building agentic capabilities into the AMS. The carriers are building on their side. Every major platform in the vertical is moving the same direction. The work your agency does runs across every one of these systems, and soon, the agents inside them will expect to talk to yours. Submissions exchanged with the carrier's agent. Records reconciled with the AMS agent. Coordination with the core-platform agents already running the work. An agency without its own agent isn't slower at intake. It's missing from the network.

This is the differentiator hiding in plain sight. Not someday. The agencies that build their AI agent now will own the relationship when the rest of the market catches up.

The honest problem is that all of this is noisy. Every vendor is shouting "agentic" this and "AI" that, and it's nearly impossible to tell what's real, what's ready, and where to even begin. That noise is exactly what keeps good agencies frozen.

Our whole job is to quiet it. We don't ask you to boil the ocean. We start you in one clear, simple, high-value place: the front door. You feel real value almost immediately, and then we help you grow from there at a pace that makes sense for your agency.

How Crux gets you started, fast

This is exactly what Crux was built to do.

We don't hand you a generic chatbot and wish you luck. We build a branded AI agent that becomes the new front door to your agency, unmistakably yours. We put it to work on the intake, servicing, and renewal moments where it creates the most value the fastest.

And I know this world from the inside. We just left Applied Systems, and before that, TechCanary, Vertafore, and ImageRight. Between us we've spent ~60 years inside the AMS, the carrier portals, the ACORD forms, and the day-to-day workflows your team lives in. We're not technologists guessing at insurance. We're insurance people who know exactly where the value is and how to reach it fast. More on the team →

You meet your new hire. Your customers meet a better experience. Your team gets its time back. And because pricing is tied to outcomes, we only win when you do.

We built Harper to show you what's possible. The next one we build will be yours.

So, what will you name your agent?