Work: From Process to Production System

Processes in production

Documents to decisions

Claims, submissions, contracts and site reports read against your own rules. Every finding traced to the page it came from.

Forms, emails and PDFs read in one system
Every exception routed to a person
Every decision logged against the record

A loss run read line by line into fields, checked against the rules and routed to a senior underwriter

Claims intake

A claim arrives by phone, email or form. The system reads it, checks the policy and opens the file, with the working attached.

Combines the policy and the claim into one file
Preserves every detail the claimant gave
Returns a ready-to-review file for the adjuster

The policy on record: coverage, deductible and status — Policy

The first notice of loss taken from a phone call, with the photos attached — Claim

The policy and the claim combined into one file, checked and assigned to a property adjuster

File

Quoting

Turn a single submission into a priced quote for every line. Rules, rates and referrals applied the same way, every time.

Reads the submission in any format or layout
Applies your rates, rules and referral limits
Sends anything outside appetite to a person

A broker submission read from email, forms and a schedule of values into one record

Precision review

Check contracts and policies against your own rules, clause by clause, with every finding traced to the page.

Targets the clauses that matter without missing the rest
Flags every change from your standard wording
Switches between documents and records on demand

A contract read clause by clause against the playbook

Reconciliation

Match every entry to its record across systems every morning. Drag the slider to compare the ledger with the reconciled view.

Ledgers, bank files, invoices and policy records
Side-by-side compare with the source system
One source of truth, every system aligned

The general ledger matched to the subledger, with the one exception routed to a personLedger

The general ledger before reconciliation, with the unmatched entries highlightedUnmatched

Blueprints into systems

Turn the written blueprint into a working system, one process at a time.

Go from blueprint to production under your own controls
Review every connection before it goes live
Bridge the business and engineering with one written plan

The written blueprint for property renewals, signed off, as a process map

Ready to build?

Start with one process and a scoping session with the people who do the work.

Talk to Us Read the Method

renewals.py

import lyrion_connectors

system = lyrion_connectors.System("renewals")
run = system.process(
    record="policy/48213",
    rules="underwriting_rules.yaml",
    approvals="head_of_underwriting",
    audit=True,
    environment="your_vpc",
)
print(run.audit_trail.url)

run.sh

# 1. Start a run
curl -X POST https://connectors.internal/v1/runs \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $CONNECTOR_TOKEN" \
  -d '{
    "system": "renewals",
    "record": "policy/48213"
  }'

# 2. Read the audit trail with the returned run_id
curl -X GET "https://connectors.internal/v1/runs/RUN_ID" \
  -H "Authorization: Bearer $CONNECTOR_TOKEN"

run.mjs

import { Connectors } from "@lyrion/connectors";

const system = new Connectors({
  token: process.env.CONNECTOR_TOKEN,
  baseURL: "https://connectors.internal/v1",
});

// 1. Start a run
const { run_id } = await system.post("/runs", {
  body: { system: "renewals", record: "policy/48213" },
});

// 2. Wait for the audit trail
let result;
do {
  await new Promise((r) => setTimeout(r, 5000));
  result = await system.get(`/runs/${run_id}`);
} while (result.status !== "done");

console.log(result.audit_trail.url);

renewals_poll.py

import time
from lyrion_connectors import Connectors

system = Connectors(
    base_url="https://connectors.internal/v1",
    token="YOUR_CONNECTOR_TOKEN",
)

# 1. Start a run
response = system.post(
    "/runs",
    body={"system": "renewals", "record": "policy/48213"},
)
run_id = response["run_id"]

# 2. Wait for the audit trail
while True:
    result = system.get(f"/runs/{run_id}")
    if result["status"] == "done":
        print(result["audit_trail"]["url"])
        break
    time.sleep(5)

Powerful without compromise

How one process compares, done four ways: accuracy, effort and turnaround.

One renewal process, by approach, illustrative
Approach	Rank	Effort	Turnaround
Lyrion Work	1
Offshore team	4
Spreadsheet macros	5
RPA scripts	9
Generic chatbot	12

Shown: one renewal process, by approach, illustrative

“The morning exposure figure used to be assembled by hand from three systems before the underwriting desk opened. Now it is waiting for them.”

Head of Underwriting, a multinational insurer

Common questions

About the work, the method and how engagements run. For a deeper dive, read the method.

What kind of work does Lyrion take on?

Processes that run on many systems and many people: intake, quoting, document review, reconciliation, renewals and reporting. We start with one process the business already runs and build a system that does that work inside your own platforms.

See the processes

Which industries does Lyrion work in?

Insurance, construction, energy and financial services today. The method is the same in every one: scope, blueprint, build and then operate, with the same engineers from the first session to the system in production.

See the industries

How are engagements structured?

Every engagement starts with a scoping session and a written blueprint. The first system covers one process and runs on your own platforms, with the next process scoped once the first is in production. We stay on to operate what we built.

See engagements

Where does the system run?

Inside your own environment: your cloud, your network, your identity provider. Nothing is copied out, and nothing is uploaded to us. Your engineers review every connection before it goes live.

Read about where it runs

Can people still make the decisions?

Yes. Anything outside the rules goes to a person, never decided for them. Approvals sit wherever your team wants them, and every override is logged against the record it changed.

See oversight

How do we get started?

Book a scoping session at lyrion.io, then we map one process with the people who do the work. You get a written blueprint before anything is built, and you decide what goes into the first system.

Book a scoping session

What happens when something breaks?

It is ours until it is fixed. The engineers who built the system operate it. Read how we operate systems. If you want to talk to them first, contact us and we will set up a call.

Is every decision audited?

Yes. Every decision the system takes is logged against the record it touched, with the rule that applied and the person who approved it.

See the audit trail

What is a written blueprint?

A document that describes the process as it really runs: the systems involved, the rules, the exceptions and who decides what. It is written with the people who do the work and signed off before anything is built.

Read about the blueprint

Can one system cover more than one team?

Yes. Once the first process runs, the next one is scoped with the team that owns it, and context carries from one department to the next.

Read the field note

Can the work scale across the business?

Each process is added one at a time, with its own blueprint and its own approvals. Engagements grow from one process to a department and then to the entire operation, without starting over.

See engagements

Can we change the rules after go-live?

Yes. Rules, thresholds and approvals are yours to change. Every change goes through change control and is logged, so you can always see what the system did and why.

See change control

Does it work with our existing systems?

Lyrion builds inside the systems you already run: policy administration, ERP, CRM, document stores and email. Your engineers review every connection.

Lyrion Work