hyperexponential (hx) is built for specialty, commercial and reinsurance carriers whose actuaries want to write their own pricing models in Python and use them inside an underwriting workflow. Swallow is built for insurers and MGAs that need a rating engine for live quoting, with models underwriters can check without reading code and AI in the build. If your risks are complex, priced one by one by underwriters, and your team wants to code its own models, hx deserves a close look. If you need a product quoting live over an API, at volume, with quick and controlled rate changes, look at Swallow.
hyperexponential in brief
hyperexponential is a software company based in London. hx has described its pricing platform, hx Renew, as a "Pricing Decision Intelligence" platform. Actuaries build models in Python on the platform, often by refactoring Excel raters, and underwriters use them to price individual risks. In a 2023 case study, Aviva's Global Corporate and Specialty team built 20 pricing models in nine months, replacing 20 separate Excel rating tools.
hx now describes itself more broadly as "the agentic platform for commercial P&C insurance underwriting". Its site covers submission triage, risk assessment, renewals and portfolio monitoring across reinsurance, US admitted, specialty and commercial lines, and MGAs. hx says its platform supports underwriting decisions on more than 75bn of commercial premium a year, and names Aviva, Allianz, Beazley, Markel, HDI and Convex among its customers.
On AI, hx lists AI-assisted model development, an Actuarial Agent, and hyperoperator, an agent that coordinates specialist agents for submission and monitoring work. hx says its agents act only within set authorities, permissions and review points, with an audit trail of what they do. It partners with Akur8, Duck Creek and Guidewire.
On 27 October 2026 Guidewire signed a definitive agreement to buy hyperexponential for 1.7bn in cash, plus up to 400m in contingent payments. Guidewire expects the deal to close by the end of January 2027 and hasn't said how hx products will be sold after that.
hx doesn't publish its prices.
Swallow in brief
Swallow is an AI-powered pricing platform and rating engine for insurers and MGAs. A model is a plain JSON file of steps that run in order: inputs, rate table lookups, calculations, transforms, exclusions, refers, excesses and endorsements. Each version carries its own tests and moves from draft to approved to published, with every step logged. One engine prices quotes over a real-time API, through embedded forms and in voice and chat, and is built to scale to millions of quotes a day. Models cover lines from pet and fleet to property, cyber and US filed programmes.
At a glance
| Swallow | hyperexponential (hx Renew) | |
|---|---|---|
| What it is | AI-powered pricing platform and rating engine | Pricing and underwriting platform for commercial P&C |
| Typical buyer | Insurers and MGAs of any size | Typically specialty, commercial and reinsurance carriers and MGAs |
| Main job | Live quoting at volume, across lines | Underwriters pricing complex, individual risks |
| Where models come from | Rating guides, Excel raters, US rate filings or any modelling tool | Written by your actuaries in Python, often from Excel raters |
| Model format | Readable JSON of ordered steps, built in code or on a no-code canvas | Python code |
| Testing | Tests on every version, expected against actual | Parallel-run validation and version control described; detail not published |
| Approvals and audit | Draft, approved, published; approvals by role; every quote records its version | Audit trail of decisions and model changes |
| AI | AI first drafts and tests; MCP server for Claude, ChatGPT and Cursor | AI-assisted model building; underwriting agents within set authorities |
| Channels | Real-time REST API, embedded forms, voice and chat | Web app, API and policy admin connections; Guidewire and Duck Creek partnerships |
| Commercial model | Talk to us | Not published |
Where they differ
What kind of risk is being priced
hx grew up in specialty and London Market business, where an underwriter prices one complex risk at a time and a model often needs logic written for that risk alone. That's where the freedom of Python counts most, and where hx is strongest. Swallow is made for rules-based rating that runs on every quote: personal lines, SME, commercial packages and specialty products sold through an API, a broker portal or an aggregator.
Python versus steps you can read
Python can express almost anything, which suits actuaries who code. It also means every change is reviewed as code, and underwriters or auditors usually need someone technical to walk them through it. A Swallow model is configuration that anyone can follow step by step: inputs with plain labels, rate tables as rows, formulas written out. Each test is a full set of inputs with its expected result, and a debug trace shows every input, lookup and formula behind any quote. Endorsement steps choose the forms and clauses for the policy documents, so wording and price come from the same rules.
Who builds and maintains the models
On hx, your actuaries usually write and maintain the Python. On Swallow your team can build models on the drag-and-drop canvas, edit the JSON, or work with an AI client. If you'd rather hand it over, send us your rating guide, spreadsheet or filing and we'll talk you through getting it live.
What AI is pointed at
hx is heading towards agents that work across the underwriting process: reading submissions, chasing missing data and watching the book. Swallow uses AI on the production model itself: a first draft from a description and spreadsheet, tests generated from a prompt, and governance reports and version summaries for reviewers. Over the MCP server, Claude, ChatGPT or Cursor can learn the format, draft a model, validate it and run its tests. They can't create, edit or publish a project. A person uploads the model to a draft, approves it and publishes it.
Cost and ownership
hx doesn't publish pricing, and the Guidewire deal raises a question worth putting to any vendor: what happens to the contract, roadmap and integrations under a new owner. For Swallow pricing, including the self-hosted enterprise offering, talk to us.
Pick hyperexponential if
- You write specialty, London Market or reinsurance business that underwriters price risk by risk.
- Your actuaries code in Python and want complete freedom in how models are written.
- You want pricing inside a wider underwriting workflow, with submission triage and portfolio monitoring.
- You run Guidewire and want a pricing platform that's joining the same group.
Pick Swallow if
- You need high-volume, real-time quoting through an API, embedded forms or an aggregator.
- You want models that underwriters, compliance and auditors can read without reading code.
- Your rates come from a rating guide, spreadsheet or filing, and you want tests on every version.
- You want general AI clients helping to build models, with people approving and publishing every release.
Next steps
Walk through a complete Swallow model on the developer page, see how quotes are priced through the API, or get in touch. For the wider market, read our comparison of Swallow and other insurance pricing platforms, or see how Swallow compares with Guidewire PricingCenter.
Based on publicly available information as of November 2026. Vendor offerings change and vary by customer, so confirm current details with each vendor.
