Qwilr

A page builder of long standing with generative features added at the top of the ladder. What the product does is deterministic and has been since 2014: a page renders from a template, a quote block computes a price from selected quantities and options, a signature is captured, a payment routes to a processor, analytics record what a viewer looked at, and a customer record system populates fields.

Remove every model and all of it continues to work, which is what the several thousand organisations using it are buying. The generative layer is real but narrow and reserved: a proposal engine that assembles personalised documents from rules the customer defines, a prefill capability, and a proposal creator promoted in the navigation, with the first two available only on the highest of three plans.

A capability sold to the top tier alone is by construction not the foundation of a product the other two tiers still use. The engine's own description points the same way, since it generates proposals built with the customer's rules, which is templating with conditional logic rather than open generation. Ask which capabilities depend on a model, and what the proposal engine evaluates when it applies a rule.

Good control over who may touch a page and nothing that checks a page before it goes out. The permission apparatus is real: edit or view only access per user, team level permissions that share pages and templates to private groups, collaborative editing, and template settings that push defaults such as expiry dates and access rules down to every page created from a template, which is governance applied at the source rather than page by page. Brand editing does the same for appearance.

What is absent from every tier in the published comparison is an approval step. No approval workflow, no discount or deal size threshold that routes a document to a manager, and no locking of individual elements against editing appears anywhere in the plan comparison, so a representative with edit access can alter a price or a term and send it without a second pair of eyes.

For a product whose output is a priced commercial offer that a buyer can accept and pay for in the same session, that is the control a reader would expect to find. The generative proposal engine at the top tier compounds the question, since it assembles documents automatically and nothing describes what reviews its output. Ask whether any approval step exists before a page is shared, and what reviews generated proposals.

Three generative capabilities named and nothing published about any of them. A proposal engine that generates personalised documents from customer defined rules and a prefill capability both sit on the highest plan, and a proposal creator is promoted in the site navigation. No model provider, family or version is named for any of the three, no model card exists, and no accuracy, quality or evaluation figure is published for generated document content.

Two questions are specific to this product rather than generic. The proposal engine is described as building with the customer's rules, and nothing states what happens when the rules underdetermine an outcome or conflict, which is the point at which a model rather than a rule decides what a buyer is offered. And prefill implies drawing on something already held, with nothing stating whether that is the customer's own prior documents, their template library, or a broader corpus.

The output here is a priced offer a buyer can accept in one click, so an error in generated content is a commercial commitment rather than a draft to correct. Ask which model powers the proposal engine and prefill, what prefill draws on, and what evaluation covers generated content before a page is shared.

Recognisable customers, a dated and counted review figure, and outcome claims that stay at the level of the anecdote. The vendor states it serves more than five thousand organisations and names several that a reader will recognise across software, media, security and higher education, which is a stronger customer roster than most products at this price point can show.

Its review citation is better practice than most: four point six out of five from nine hundred and twenty two reviews, with the count given rather than the score alone, published on the pricing page where a buyer is deciding. The outcome evidence is thinner.

The strongest published claim comes from a customer rather than the vendor and does carry internal arithmetic, describing an hour saved per deal closed across more than two thousand closed deals for more than two thousand hours saved, which at least ties the saving to a countable event.

Everything else is a percentage without a base: increases in win rate, reductions in time to close and in time spent creating materials appear as headline figures with no population, period or method attached. Ask what population and period the win rate and cycle time figures cover, and how time saved was measured.

Recipients are counterparties in an active deal, so this axis applies mostly by scope, and one published capability complicates that. There is no campaign tool, no sequence, no contact database and no bulk dispatch anywhere in the product. What leaves the platform is a page link sent to a named buyer and notification correspondence around it, which is transactional.

The complication is that pages are shareable links, and the vendor markets the identification of viewers who were not the original recipient, meaning the product anticipates and reports onward internal forwarding. Those secondary viewers never received anything from the seller, and nothing published addresses what they are told or whether the domain restriction and identity verification controls are expected to be used to manage that.

Notification correspondence is also unaddressed: nothing states what a buyer receives automatically, whether they can stop it, or what sender identity appears once the customer has enabled a custom domain and removed vendor branding from emails. Ask what correspondence a recipient receives automatically and whether they can decline it, and what a forwarded viewer is told.

One retention position published properly and the rest of the instruments out of reach. Analytics history retention is stated as a plan feature at sixty days, one hundred and twenty days and unlimited across the three tiers, which is an explicit retention ladder for the most sensitive data class the platform holds. Publishing a retention period at all is unusual; publishing it as a comparison row is clearer still.

It also means how long records about document viewers are kept is set by what the customer pays rather than by a uniform policy, which a reader should weigh both ways. Beyond that little was established. A privacy policy, terms and a dedicated trust and security page are all published and the last was not retrieved on this pass, so no processing agreement, subprocessor list, transfer mechanism or data protection contact could be confirmed.

The corpus deserves naming: executed agreements and their commercial terms, payment records, and identity level analytics on the individuals who read each page, including people the document was forwarded to who have no relationship with the seller. The vendor is an Australian company selling globally, so its home privacy regime is not the one most comparable material addresses. Ask for the processing agreement and subprocessor list, the retention period for documents and payment records, and what a page viewer is told about identity analytics.

Almost nothing on this axis applies and the grade records the scope rather than a shortfall. The vendor supplies no contact database, no firmographic data, no intent feed and no enrichment. Page content comes from the customer's own template library, their saved content blocks, their pricing and fields pulled from their own connected customer record system, so there is no purchased corpus whose origin a buyer needs to trace and no third party data licence sitting underneath the product.

That is a favourable position worth stating plainly. Two smaller questions remain open. The vendor publishes a library of ready made templates spanning many use cases, and nothing states who authored them or whether any professional review sits behind commercial or contractual language a customer may send unchanged.

And the prefill and proposal generation capabilities produce text whose basis is undescribed, which matters more here than for marketing copy because the output is an offer a buyer can accept with one click. Ask who authors and reviews the template library, and what the generative features draw on.

Every connection runs through a sanctioned interface and the vendor publishes one of its own. Named integrations cover five customer record platforms, an automation service, a messaging tool, an accounting package that invoices automatically when a page is accepted, and a payment processor, each through that product's own documented route.

Developer documentation for the vendor's own interface sits on a dedicated subdomain and access is included on every plan, so other software builds on this product through a published path as well. Nothing scrapes a professional network, no browser extension pulls data off one, and no social automation exists anywhere, so the exposures that most often surface on this axis do not arise. Held below the top band on two points.

No conformance position is stated in the vendor's own words for any connected platform, and nothing describes what happens if a connector's terms change. And the payment capability is a layer the vendor operates on top of a third party processor, taking its own percentage on each transaction at the entry tier, which means the customer's payment relationship runs through an arrangement the vendor intermediates and nothing published describes what happens to collected funds or stored payment credentials if that arrangement ends. Ask what governs the payment layer if the processor relationship changes, and for the stated position on each connected platform.

Qwilr's trust page says neither Qwilr nor the large language models it uses train AI models on customer content or personal data. Its trust dashboard names Groq as the AI inference platform, processing in the United States without storing data there, customer data is stored and processed in Sydney, and all customer data is deleted within 30 days of a subscription being cancelled.

The Smart Proposal Engine lets leaders choose where AI may write and keep it away from pricing and contract sections, and the MCP connection on every plan runs under the same terms. How long prompts and generated text are kept during a subscription is not stated.

Two features here work against the person receiving the document and neither is accompanied by any statement about what they are told. The first is identity analytics, sold on seeing exactly who is viewing a page, together with marketing that promotes identifying buyers the seller did not know about.

Because a page is a shareable link, those additional viewers are typically colleagues the document was forwarded to internally, who received nothing from the seller and have no relationship with them, and nothing published states what they are told about being identified. The second is more particular to this product. The deliverable is a live page rather than a file, and the vendor markets that a seller can edit it after sending and the buyer will automatically see the current version.

That means the document a buyer read when they formed their view may not be the document in front of them when they sign, and nothing published states whether a recipient is notified that a page changed or can see what it previously said. An audit trail is included on every plan, which protects the seller's record, and nothing indicates the buyer has access to it.

Against those sit real mitigations: legally binding signature, and optional controls for password, expiry, view caps, domain restriction and asking a viewer to identify themselves. Ask what a forwarded viewer is told about identity tracking, and whether a buyer is notified when a page changes after they have seen it.

A documented interface available on every plan, and a connector set that is short but current. Developer documentation sits on its own subdomain and is linked from the main navigation, and interface access is included at all three tiers rather than reserved for the largest buyers, which is the opposite of the usual pattern and lets a small team automate from the start.

Named integrations cover five customer record platforms, an automation service reaching thousands of applications, a messaging tool for engagement alerts, an accounting package that raises an invoice automatically when a page is accepted, and a payment processor. The accounting behaviour is worth noting because it closes the loop from acceptance to invoice without a person. Two things hold it below the top band.

Interface use carries usage costs at every tier and support for it is help documents only on the entry plan, so the capability is included while the assistance and the consumption are not. And the deeper customer record connections are tiered, with two of the five reserved for the upper plans, so a buyer on the entry tier gets the interface but not the integration they may most need. Ask what interface usage costs at the tier being quoted, and which record platform integrations are included at each tier.

Nothing on hosting, region, tenancy or recovery was reached on this pass, and a dedicated trust and security page exists that was not retrieved, so this records the limit of the pass rather than a finding of absence. The question has a particular shape for this product that a reader should understand.

The deliverable is a page served from the vendor's infrastructure to a buyer's browser at a public link, so the availability of that infrastructure is not a back office concern but the thing a prospect either can or cannot open during a live deal, and nothing published states an uptime position or a recovery objective.

The vendor is an Australian company selling into Europe, North America and its home market, and where executed agreements, payment records and viewer analytics are stored is unstated in every direction. A custom domain option lets pages be served from the customer's own subdomain, which changes what a buyer sees without changing where anything is hosted. Ask which provider and regions serve pages and store agreements, whether any regional option exists, and what the availability commitment and recovery objectives are.

Qwilr completes a SOC 2 Type 2 audit each year with controls monitored continuously in Drata, and its trust dashboard lists the 2025 SOC 2 Type 2 report beside a cyber insurance schedule, security policies, control monitoring status and its subprocessors.

The trust page adds penetration testing every year and after significant platform changes, a vulnerability disclosure program, encryption in transit and at rest, OAuth SSO on every plan, SAML SSO on Scale and notice to account administrators when an incident is suspected. Payments run through Stripe. Page level controls protect the shared document itself: passwords, expiry dates, view caps, domain restricted access and viewer identity checks. The audit period and the terms for opening the report are not stated on the trust page.

Qwilr publishes the whole price list. Starter is $35 per user a month on annual billing or $49 monthly; Growth is $55 per user a month with a five user minimum and Scale $75 per user a month with a ten user minimum, both billed annually only, and extra users cost the plan rate.

Document automation is priced in full: 40, 200 and 600 documents included on the three plans (none on Starter billed monthly), pay as you go at $2.50 a document, and credit bundles from 250 documents at $2 each ($500) to 5,000 at $1.50 each ($7,500) on Scale. QwilrPay takes 0.09% a transaction on Starter and nothing above it, analytics retention and seat reduction rules are stated, a 14 day trial needs no card, and what happens to pages at trial end and at cancellation is set out. Onboarding packages and professional services are the only unpriced items.

Two export mechanisms on every plan, and a cancellation term that switches off the thing the customer's buyers are holding. Document export to a portable file and a complete audit trail are both included at all three tiers, so a customer can retrieve their agreements and the record of what happened to them without asking, and offering both at the entry tier rather than reserving them is worth crediting.

The termination position is where this becomes serious, and it is unusually explicit for something so consequential. The vendor states that on cancellation the pages are turned off and account access ends. Because the deliverable is a live link rather than a file, that means every proposal a customer has sent, including any a buyer is currently reading or has bookmarked to sign, stops rendering.

The contrast with the trial term is instructive: a lapsed trial leaves pages live for thirty days, while cancellation states no grace period at all. Nothing published covers bulk export of templates, saved content blocks, brand settings or analytics history, which together represent the accumulated configuration work. Ask whether any grace period applies to live pages after cancellation, whether templates and saved blocks export, and what becomes of analytics history and payment records.

The product sends notification correspondence on every plan and publishes nothing about how. Page links, view alerts, acceptance confirmations and signature requests all travel by electronic mail, and a confirmation that fails to arrive is a signature that does not happen.

Nothing published describes which infrastructure sends, whether correspondence leaves from vendor domains or the customer's own, what bounce or complaint handling applies, or whether a delivery failure is surfaced to the sender at all. Two published capabilities sharpen the question rather than answering it.

A custom domain lets pages be served from the customer's own subdomain, and vendor badges can be removed from confirmation emails, so at the upper tiers the correspondence can appear entirely as the customer's own, and nothing states who then configures the sender authentication records or carries the consequence when they are wrong. Volume is low and the correspondence is transactional, which keeps the exposure well below that of a campaign product. Ask which infrastructure sends notifications, whether custom domains extend to correspondence, and who configures authentication when vendor branding is removed.

Coverage is stated across three dimensions at once and the seat ladder makes the boundaries concrete. Industries are named as software and technology, marketing and digital agencies, event services, real estate and universities, which is an unusually varied set and includes one, higher education donor reporting, that is a genuinely different motion from selling. Roles run beyond sales to marketing, customer success, revenue operations, recruitment and developers.

Use cases extend past proposals to contracts, digital sales rooms, mutual action plans and customer onboarding, so the product is positioned across the deal rather than at one point in it. The seat minimums of one, five and ten state the team size each tier assumes without a buyer having to infer it, and student and non profit discounts are offered on request. Currency coverage is handled properly with two currencies, a toggle and a tax note. Two things hold it below the top band.

Pricing exists in Australian dollars for one market and United States dollars for every other, so a buyer in Europe or elsewhere transacts in a currency that is not theirs and nothing addresses that. And no guidance is given on deal volume or value at which the product pays back, which for a per user tool with a per document charge on top is the calculation that decides fit. Ask what document volume and deal value the tiers assume, and whether local currency billing exists beyond the two published.