CopyCat AI
CopyCat AI is an AI workspace for insurance brokers that turns carrier quote PDFs into branded proposals, cited quote comparisons and renewal risk reports, plus live-formula employee benefits workbooks, with every figure traced back to its source page.
What is CopyCat AI?
CopyCat AI is an AI workspace for insurance brokers, published by CopyCat Technologies, Inc., a Y Combinator Winter 2025 company listed in Seattle. Both founders, Abhi Balijepalli and Graham Sabin, sat the commercial broker exam before building it, and the first version was assembled in person alongside working brokers.
The target is the document work sitting between a carrier and a client. Brokers stitch together carrier portals, ACORD forms, AMS exports and quote documents by hand; CopyCat automates that step without asking the agency to change anything else.
Three modules share one intake. Proposal Generation turns carrier quote PDFs into a branded, editable proposal across commercial and personal lines. Risk Analysis normalises competing quotes into a side-by-side table, answers questions in chat across every uploaded page, compares a renewal against the expiring policy, and exports either a one-page decision brief or a full report. Benefits converts new quotes, the incumbent plan and an employee census into a live-formula Excel workbook.
The intake is deliberately wide: carrier PDFs, ACORD forms, expiring policies, loss runs, fleet and driver schedules, statements of value and census spreadsheets, with no page cap. The publisher claims 200+ pages per document and 1,000+ carrier formats normalised without per-carrier setup. Coverage spans Property, General Liability, Workers' Comp, Cyber, BOP, Umbrella, Management Liability, E&O, Crime, Inland Marine and Flood, plus E&S markets, Lloyd's syndicates, MGAs and program business, alongside personal lines.
The governing principle is traceability rather than free generation: every extracted value opens the exact page of the carrier document it came from, and the vendor states nothing is invented. Output is edited in a word-processor-style editor, versioned, then exported as a branded PDF.
A shared workspace makes every document visible across the organisation from creation, with owner avatars and a timestamped audit log that doubles as an E&O paper trail. Security claims include SOC 2 Type II with the report available under NDA, AES-256 at rest, TLS 1.3 in transit, multi-factor authentication, role-based access and HIPAA compliance with a Business Associate Agreement.
Self-declared results: 2,000+ proposals generated, seven to ten hours saved per broker per week, and an error rate below 0.03%. Access is by booked demo only, with no self-serve signup and no published price.
What it does
- Turn a set of carrier quote PDFs into a branded, editable insurance proposal for commercial or personal lines
- Compare competing carrier quotes side by side on premium, limits, deductibles, valuation and exclusions, with the gaps flagged
- Ask questions in chat across every uploaded document and get answers pinned to the exact source page
- Compare a renewal against the expiring policy and surface every material change in coverage, deductible or premium
- Export a one-page client decision brief or the full editable risk analysis from the same cited conversation
- Convert benefits quotes, the incumbent plan and an employee census into a live-formula Excel workbook
- Share every document across the agency in one workspace with a timestamped, exportable audit log
When to use CopyCat AI / When not to
A quick filter to help you decide if CopyCat AI is the right fit.
When to use CopyCat AI
- Commercial lines producers and account executives in US independent agencies, who spend hours rebuilding a proposal from carrier PDFs for every account
- Account managers and client service representatives handling renewals, who need this year's quote compared against the expiring policy with every material change flagged
- Employee benefits brokers who assemble market summaries, contribution sheets and census-driven workbooks by hand each benefits season
- Agency operations and practice leaders who need an automatic, timestamped audit trail behind every client document for E&O purposes
- Brokerages that refuse to abandon their own proposal template and want their existing Word, PowerPoint or InDesign format reproduced rather than replaced
When not to use CopyCat AI
- Brokers outside the United States: the coverage vocabulary, forms and market logic are built for US commercial lines, and the vendor itself points markets that expect a Statement of Advice or a Product Disclosure Statement towards a competitor
- Anyone who wants to test software the same afternoon without talking to a salesperson: there is no self-serve signup, no free plan and no free trial, only a booked demo
- Buyers who need a published price before they will engage, since no amount, tier or currency appears anywhere on the site
- Development teams looking for an API or a mobile app: no API documentation exists on the site and there is no iOS or Android application
- Agencies expecting a system of record or a deep AMS/CRM integration, since the product deliberately sits alongside those tools and connects to none of them
How to use CopyCat AI
A typical end-to-end flow, from setup to results.
- Book a demo from the site: there is no self-serve signup, so the booked call is the only way in
- Bring your own material to that call — the vendor suggests the carrier PDF that normally breaks things and the template with the awkward tables
- During onboarding, hand over your existing proposal format in Word, PowerPoint or InDesign, or pick one of the 30+ library themes
- Let CopyCat rebuild the template for you: roughly 80% is standard and the remaining 20% is configured by the vendor, so your team never edits a config file
- Sign in to the web app at app.runcopycat.com through Clerk, with single sign-on available
- Upload a single carrier quote or a full multi-carrier, multi-line account, in PDF, Word, Excel or email attachments
- Review the generated proposal, where coverage tables, premiums, exclusions, conditions and your standard language are already placed
- Click any citation to open the exact page of the source carrier PDF beside the document, then edit, save a version and export the branded PDF
- For a renewal or a competing quote set, open Risk Analysis, ask your question in chat and export the one-page brief or the full cited report
- For employee benefits, upload the new quotes, the incumbent plan, the census and the contribution rules, then download the live-formula Excel workbook
Pros & Cons
Pros
- Traceability is designed in rather than promised: every figure opens the page of the carrier PDF it came from, which is what makes the output defensible in front of a client
- Your existing proposal format is reproduced rather than replaced, and the vendor does the template build so no one on your team touches a configuration file
- Nothing to migrate and nothing to integrate: the tool sits alongside your current AMS, CRM, carrier portals and email
- A contractual guarantee, not just a policy line, that customer data is never used to train, fine-tune or improve any model
- Unusually mature security for a company this young: SOC 2 Type II, HIPAA compliance with a BAA, AES-256 at rest and TLS 1.3 in transit
- Commercially straightforward on paper: a flat monthly fee with no annual contract, no lock-in, no implementation charge, no per-module add-on and no per-document metering
- One upload feeds three deliverables — the proposal, the cited comparison and the benefits workbook — and the audit log writes the E&O trail automatically
Cons
- No price is published anywhere: there is no pricing page at all, and no amount, tier or currency appears in the FAQ or in any of the five comparison pages
- No self-serve entry, no free plan and no free trial — the only route in is a booked demo, and the vendor's own equivalent is described as guided onboarding
- No documented API despite the Terms referring twice to APIs and API keys, and no iOS or Android application
- No postal address is published on any page: the only identity disclosed is the legal name in the footer copyright and in the Terms
- A single generic mailbox serves legal, GDPR and commercial enquiries, and the site's own structured data labels it as a sales address
- Data protection is thin around the edges: no DPA, no published subprocessor list, no Article 27 EU representative and no disclosed hosting country
- The headline numbers — 2,000+ proposals, an error rate below 0.03%, seven to ten hours saved weekly — are self-declared with no published method, and the SOC 2 report is only released under NDA
Pricing & Plans
There is no permanent free plan and no free trial. Pricing is described only as a flat monthly fee, with no annual contract, no lock-in, no implementation charge, no per-module add-on and no per-document metering; new capabilities are said to arrive within the plan already paid for. No amount, currency or tier is published anywhere on the site, and there is no pricing page. Commercial terms are set individually in an order form, invoiced in advance and payable within thirty days, with fees revisable on thirty days' notice before renewal. A figure can therefore only be obtained by booking a demo.
- A single flat monthly subscription covering the whole product — no named tiers
- no seat
- module or document-based levels
- and no published amount
- commercial terms are agreed individually in an order form
Data, GDPR & hosting
A consolidated view of how CopyCat AI handles your data.
GDPR overview
GDPR is mentioned, but cautiously. The privacy policy and the Terms both state, in identical wording, that the publisher processes data in accordance with relevant data protection regulations "including CCPA and GDPR", prefaced by the reservation "where applicable". Do Not Track signals are honoured where the law requires it. The rights listed are the expected ones and are presented as depending on your location: access, correction, erasure, restriction or objection, portability in a structured format, and withdrawal of consent, with a stated thirty-day response time. All requests go to a single generic address. Beyond that, the implementation is thin: no Data Processing Agreement is published or offered, no Article 27 EU representative is designated, no Data Protection Officer is named, no legal basis for processing is stated, no hosting country is disclosed and no European entity exists. The privacy policy is dated June 2026.
Who owns the data?
The customer keeps ownership. Section 5 of the Terms states that you retain all rights, title and interest in the data you provide or process through the service, and that CopyCat AI claims no ownership of that Customer Data, which covers policy documents, carrier quotes and client information. The vendor takes only a limited, non-exclusive, worldwide licence to access, use and process it as far as necessary to provide, maintain and improve the service. Sharing is confined to service providers under confidentiality, professional advisers, law enforcement on legal process, and a possible business transfer with notice. The publisher states it does not sell personal information or Customer Data.
Reuse rights
The customer is free to reuse everything the tool produces: proposals, comparison reports and benefits workbooks are exported as branded PDF or Excel files for the agency's own client work, and no permission step, watermark, attribution requirement or output-licensing clause appears in the Terms. The vendor's own use of Customer Data is narrower than usual for an AI product. Section 5 carries an explicit contractual guarantee that policy documents, carrier quotes, client information and anything else processed through the service will not be used to train, fine-tune or improve any generative AI or large language model, and will never be aggregated into training datasets or sold to third parties. The privacy policy repeats the point, and the product interface displays the same statement under its chat. Processing is limited to delivering the requested service: extraction, analysis and document rendering. Third-party processors are barely disclosed: only Clerk, for credentials and single sign-on, and PostHog, for analytics, are named anywhere.
Data retention & training
Hosting summary
No hosting country or region is disclosed. Neither the privacy policy nor the Terms names a data centre, a cloud provider or a jurisdiction of storage, and there is no dedicated trust or security page — the security content amounts to one homepage section. No subprocessor list is published either; the only third parties named anywhere are Clerk, for credentials, API keys and single sign-on, and PostHog, for analytics, both mentioned in passing. The privacy policy otherwise refers to unnamed "service providers" for hosting, analytics and communication. What is documented is the protective measures rather than the location: encryption in transit via TLS and at rest via AES-256, role-based access control, multi-factor authentication, audit logging and continuous monitoring, regular security assessments and penetration testing, encrypted credential vaults, claimed isolation of each agency's data, SOC 2 Type II certification with the report available under NDA, and HIPAA compliance with a Business Associate Agreement for customers handling protected health information. The website itself resolves to a US anycast node, but that describes the marketing site, not where client documents are stored.
Things to keep in mind
Risks and trade-offs to weigh before adopting CopyCat AI.
- Automation bias is the central risk: output that looks finished and carries citations invites a quick skim instead of a real review. The interface itself warns that the tool can make mistakes, and a wrong limit or a missed exclusion in a bound proposal is an errors-and-omissions claim, not a typo
- Skill erosion over time: junior producers who never assemble a proposal by hand may never learn to read a carrier form closely, and the ability to spot a buried clause is exactly what the tool is being trusted to replace
- You cannot budget or compare offers before engaging, since no price, tier or currency is published anywhere and the only route to a figure is a sales conversation
- The performance claims — 2,000+ proposals, an error rate below 0.03%, seven to ten hours saved per broker per week, 1,000+ carrier formats — are self-declared with no published methodology, and the SOC 2 Type II report is only released under NDA, so neither can be checked before signing
- The GDPR position is a careful "where applicable" formula rather than a commitment: no Data Processing Agreement, no Article 27 representative, no named DPO, no stated legal basis and no disclosed hosting country, which matters if you handle any European client data
- Corporate transparency is thin: no postal address on any page, one generic mailbox for legal, privacy and sales alike, and testimonials that are mostly first-name-and-role anonymous
- Commercial exposure sits in an individually negotiated order form: fees are non-refundable, revisable on thirty days' notice, and late payment accrues interest at up to 1.5% per month
Setup & Integrations
Technical difficulty
Very low for the customer, because the work is deliberately shifted to the vendor. No technical skill is required and no one on your team edits a configuration file: the publisher rebuilds your proposal template during onboarding, announced in days rather than weeks, with one customer reporting four days from start to sending real proposals. Nothing is migrated and nothing is integrated. Day-to-day use is uploading PDFs, reviewing, editing in a word-processor-style editor and exporting. The real friction is organisational: you must book a demo to get access, then gather your existing template and, for benefits, your census and contribution rules.
Deployment
Supported languages
Behind CopyCat AI
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement CopyCat AI.
Frequently asked questions
What does CopyCat AI actually produce?
How much does it cost?
Is there a free trial or a free plan?
Can it invent coverage or hallucinate figures?
Is customer data used to train AI models?
Do we have to change our existing systems?
Can our agency keep its own proposal template?
What happens to our documents if we leave?
Should you pick CopyCat AI?
CopyCat AI is a narrow tool, and that is its strength. It serves one trade, US insurance broking, and one moment of that trade: the document that goes to the client. Within those limits it is unusually well built. The product bets on traceability rather than free generation — every limit, deductible and premium opens the page of the carrier PDF it came from — which is precisely what makes an AI-produced proposal defensible when a client or an errors-and-omissions reviewer asks where a number came from. Reproducing an agency's own template instead of imposing a vendor layout is the second good decision, and the vendor doing the configuration work rather than the customer is the third.
The security posture is more mature than the company's age suggests: SOC 2 Type II, HIPAA with a Business Associate Agreement, and a contractual, not merely editorial, guarantee that client policy data never trains a model.
The reservations are commercial and evidentiary rather than technical. Nothing about the price is public, there is no way in except a booked demo, and the impressive figures — two thousand proposals, an error rate below 0.03%, seven to ten hours saved weekly — are self-declared with no published method. No postal address appears anywhere, a single generic mailbox handles legal and privacy matters, and the data-protection paperwork stops well short of what a European buyer would expect.
This is a young company, its domain barely eighteen months old, with two documented customer cases and a specialised bet. For a US brokerage that already knows the proposal step is eating its week, the vendor's own suggestion is the right test: bring the worst carrier PDF you have and your own template, and judge the output.
- Choosing a selection results in a full page refresh.
- Opens in a new window.