How We Work

Engagement model

We work two ways: project engagements, where our engineers build and run one agent from the services list, and staff augmentation, where a senior engineer embeds with your team on AMS integration and agent work. Both are billed hourly, on a time-and-materials basis, invoiced as the work is done.

Scope comes out of an initial conversation and a written plan, and it is revisited as the work teaches us both something. We do not publish rates, and nothing on this site is a price. What you always have is visibility: every invoice itemizes the hours, and the engineer who bills an hour is the engineer who worked it. After a build you can stop, run the agent yourself with the runbook we hand over, or keep us on hourly operations work as carriers, portals, and your AMS change.

How a build runs

1. Scoping call. Which AMS, which workflow, which carriers, which mailboxes. You leave with a written plan and a start date, or with a straight answer that your setup is not a good fit yet.
2. Access. We send an exact list of scoped credentials: AMS API access (Applied Epic SDK or API program, HawkSoft Partner API, EZLynx API), mailbox delegation, carrier portal logins. Read-only where possible; write access only to the activity log and only when we switch it on.
3. Silent run. The agent runs against your live data for one to two weeks producing output nobody acts on. We compare it to what your team actually did and tune from the differences. This is the step most vendors skip, and it is why our agents are right on your book instead of on a demo book.
4. Supervised go-live. One reviewer, every action approved, daily check-in.
5. Rollout. The whole team, a runbook, and an accuracy report with the numbers from the silent run and the supervised week.

What you receive

• The running agent, deployed in an AWS account dedicated to your agency.
• Source code and infrastructure definitions. You own them; there is no license fee and no lock-in.
• A runbook: how it works, how to change the rules, what to do when a carrier changes its portal.
• An accuracy report against your real cases.
• Every action logged with its reasoning and the source documents it relied on.

Human in the loop, always

Our agents draft; licensed people decide. Nothing goes to a client or a carrier without an approval click, and the approval is recorded. This is not a limitation we intend to remove. In a commercial agency the E&O exposure lives in the send, and a person has to own it.

Security and compliance posture

• Single-tenant. Each agency's agent runs in its own AWS account in a US region. No shared queues, no shared databases.
• Secrets. Credentials in AWS Secrets Manager, rotated on a schedule, never present in prompts, logs, or source.
• Data handling. Client and policy data is processed to do the task and stored only as long as the task and its audit trail require. It is not used to train any model. We use model providers under terms that exclude training on customer data.
• Encryption. In transit and at rest, using AWS-managed keys unless you require your own.
• Least privilege. The agent gets the narrowest AMS permission set that does the job, and we document which.
• Audit. Every read and write is logged with a timestamp, the account touched, and the reasoning.
• Your policies. If your carriers, your E&O carrier, or a state regulator require specific controls, tell us at scoping and we build to them. We will not claim a certification we do not hold; ask us which we hold at the time you engage.

Timeline at a glance

Typical arcs, not commitments; actual pace depends on carriers, intake channels, and API access.

Who does the work

Senior US-based engineers on our own team. No offshore hand-off, no junior bench. The engineer on the scoping call is the engineer on the build.