Duck Creek Rating suits P&C carriers that run Duck Creek Policy, or plan to, and want rating, rules, forms and workflow held in one product configuration. Swallow suits insurers and MGAs that want a rating engine any policy system can call over an API, with readable models, tests on every version and AI in the workflow. If you're building on the Duck Creek suite, its rating engine sits right beside your policy and product definitions. If you aren't, or you want a product live soon, Swallow is worth a look.
About Duck Creek Rating
Duck Creek Technologies sells a suite of P&C core systems covering policy, billing, claims and rating, which it says can be "implemented independently or via a combined approach".
Duck Creek Rating is the suite's rating engine, described in Duck Creek's current product sheet as "the enterprise-grade P&C rating engine" for personal, commercial and specialty lines. Duck Creek's public material describes:
- Low-code configuration of "rates, rules, algorithms, and workflows", so business users can define products without writing code.
- A single product definition that combines rating, rules, forms, pages and data, with product inheritance to reuse structures across lines and regions.
- What-if modelling to see how a rate change would affect the book before it goes live, and to produce data for filings.
- Native support for ISO, AAIS and NCCI bureau content.
- Rate change and approval workflows, audit controls, and automated regression testing through the Duck Creek Test Automation Center.
Duck Creek Rating can run as a "black box" rating engine next to a carrier's existing policy system, or together with Duck Creek Policy. It's delivered as SaaS through Duck Creek OnDemand, where Duck Creek hosts and upgrades the software, and Duck Creek runs on Microsoft Azure.
On AI, the product sheet says Rating is "powered by an agentic platform", and Duck Creek markets a broader Agentic AI Platform. We didn't find published detail on AI features specific to rating. Duck Creek doesn't publish pricing.
About Swallow
Swallow is a pricing platform and rating engine for insurers and MGAs. Each model is a JSON file of ordered steps: inputs, rate table lookups, calculations, exclusions, refers, excesses and endorsements. Teams build models on a visual canvas, from a prompt or an uploaded spreadsheet, or with an AI client over the MCP server. Every version keeps its own tests and moves from draft to approved to published. Any system can price a quote through the REST API, and the same model also powers embedded forms and voice and chat AI.
At a glance
| Swallow | Duck Creek Rating | |
|---|---|---|
| What it is | Pricing platform and rating engine | Rating engine within the Duck Creek core suite |
| Typical buyer | Insurers and MGAs | Typically P&C carriers, often on or moving to Duck Creek Policy |
| Policy system | Any. Called over a REST API | Duck Creek Policy, or standalone with another policy system |
| Where models come from | Rating guides, spreadsheets, US rate filings or any modelling tool | Configured in Duck Creek's low-code product tools; ISO, AAIS and NCCI content |
| Model format | Readable JSON of ordered steps; rate tables upload and download as CSV | Duck Creek product definition |
| Testing | Tests saved on each version and run on a draft before approval; AI can generate tests from a prompt | What-if modelling; automated regression testing via Test Automation Center |
| Approvals and audit | Draft, approved, published; roles decide who approves; every step logged with who and when | Approval workflows and audit controls |
| AI | AI builds models and tests from prompts; MCP server for Claude, ChatGPT and Cursor | Agentic platform described; rating-specific detail not published |
| Deployment | Cloud platform run by Swallow | SaaS through Duck Creek OnDemand on Microsoft Azure |
| Commercial model | Talk to us | Not published |
Where they differ
One product definition, or an engine on its own
Duck Creek Rating's real strength is the product definition. With rating, rules, forms and pages kept in one configuration alongside the policy system, a change is made once and flows everywhere. Built-in bureau content also helps US carriers working from ISO, AAIS or NCCI. Swallow is independent of any policy system. Systems call it over the API, which suits carriers on other platforms, MGAs working on a broker's system, and direct or comparison-site quote journeys.
Who does the building
Core system rating is usually configured by trained specialists, often with an implementation partner. On Swallow, your pricing team owns the model and its releases. They can build on the canvas, start from a prompt and a spreadsheet, or have an AI client draft the JSON. If you'd like help getting started, send us your rating guide, spreadsheet or filing and we'll talk you through it.
Reading and checking a model
Both products let you test a change before it goes live. In Swallow the model is one readable file of ordered steps, and its tests live on the version. Each test is a full quote with the result you expect, so running the suite on a draft shows which prices move before anyone approves it. The debug trace for any quote shows every input, lookup, formula and rule with the values behind it, and each quote records the version that priced it. Our guide to testing insurance pricing models covers how to build a suite that catches the quiet mistakes.
Starting from filings
Duck Creek's bureau content support is aimed at carriers that adopt ISO, AAIS or NCCI material. Swallow lets you search US insurance filings and turn them into rating engines, which helps when you're moving your own filed rates onto a new engine or entering a new state. See building pricing models from US rate filings.
Where the AI sits
Duck Creek describes its platform as agentic but hasn't published what AI does inside rating. In Swallow, AI works on the model and the process around it: generating a model from a prompt, generating tests, writing version summaries and governance reports, and answering questions about your quote data. Over MCP, Claude, ChatGPT or Cursor can read the rules, draft a model, validate it and run its tests. Agents can't create, edit or publish a project. A person uploads the model to a draft and approves and publishes it in the dashboard.
Documents and channels from the same model
Endorsement steps in a Swallow model pick the schedules, forms and clauses that apply to a risk, and a template turns the quote response into policy documents. The same published model also serves embedded quote forms and voice and chat AI, so every sales channel prices from one place.
Cost
Duck Creek doesn't publish pricing, so ask for the full cost up to the first live quote, including implementation and integration with your policy system. For Swallow, talk to our team.
When Duck Creek Rating is the better choice
- You run Duck Creek Policy, or are moving to it, and want rating in the same product definition as rules and forms.
- You're a US carrier relying on ISO, AAIS or NCCI content and want it supported natively.
- You want one core system vendor for policy, billing, claims and rating.
When Swallow is the better choice
- Your policy system isn't Duck Creek, or you're an MGA without a core system programme.
- You're launching or migrating a product and want it live without waiting on a core system programme.
- Your rates already exist in a rating guide, spreadsheet or filing, or in a modelling tool you use today.
- You want models people can read and audit, with tests on every version.
- You want endorsements in the model to drive your policy documents.
- You want the same model to serve your API, quote forms and conversational channels.
Next steps
See how a Swallow model is built, tested and released on the model building page. Our overview of Swallow vs insurance pricing platforms sets core system rating modules against the other options, and the insurance rate change process shows how a change moves from draft to live.
This comparison reflects publicly available information as of November 2026. Vendor offerings change and vary by customer, so check current details with each vendor.
