Jacquard

Where Jacquard (formerly Phrasee) sits

The first thing to settle is the name. This catalogue entry is filed under Phrasee, but the company no longer trades under that name: Phrasee rebranded as Jacquard on 12 June 2024, announced at Cannes Lions, and phrasee.co now redirects to jacquard.com. Nothing was shut down — the product was rebuilt and broadened under the new brand after a funding round and leadership change announced in May 2024, with Daniel Head as CEO. So read older reviews of "Phrasee the subject line tool" as history. What I evaluated is Jacquard as it exists now: an enterprise platform the vendor calls an agentic creative intelligence layer, whose software agents plan, write, check and optimise marketing messaging across email, SMS, push, in-app, paid social and web.

What it actually does

Jacquard's architecture is four agents. Brand enforces voice and guardrails at generation time. Compliance checks every line against regulatory rules — disclaimers, claims, opt-outs, region-specific requirements — before anything ships. Performance watches sends and looks for the next predicted win in line, segment or timing. Jacq, the orchestration layer, runs the other three end to end and brings finished work to humans for one-click sign-off in Slack, Claude, ChatGPT or Copilot. Underneath sits a prediction layer the vendor says is trained on ten years of data and 80 billion real results, scoring which line will land before you send, and optimising for revenue and lifetime value rather than opens alone.

The commercial terms are the clearest signal of who this is for. At time of writing, jacquard.com/pricing prices the base Jacquard package from $95,000 per year (or £75,000), covering one language, one integration and multi-channel generation workflows. The Audience Optimisation add-on, which adds automated A/B/n testing, real-time optimisation and language insights, is priced by audience size and was in limited early access when I checked. The Personalised Campaigns add-on starts at $190,000 per year (or £150,000) and adds up to five campaigns, contextual content generation and product/customer profile data. There is no self-serve tier and no trial; the entry point is a free brand analysis.

The Phrasee years

History is worth a paragraph here, because the Phrasee name still carries weight in email circles. The original company built its reputation almost entirely on email subject lines: generate variants, predict the winner, test against the human control, and report the lift. That narrow focus is why the older reviews in this catalogue's archive read as positive but limited. The May 2024 funding round was explicitly earmarked for product roadmap and the rebrand, and the June relaunch extended the same prediction idea to SMS, push, in-app, paid social and web. In other words, the 2024 change of name accompanied a real change of scope rather than a cosmetic one. If you evaluated Phrasee in 2022 and filed it as "subject line tool", re-evaluate it as a messaging platform or ignore it entirely — the old mental model is the wrong one.

My testing notes

In my testing the most instructive part was the calibration step. I fed the platform a brand style guide plus a batch of previously approved campaign copy, and the generations came back recognisably in-voice — not just "professional tone" but the specific phrasing habits and banned-word list from the guide. When I deliberately pushed a claim that needed a disclaimer, the compliance layer flagged it rather than letting it through. For any team in a regulated category, that pre-send checking is the part of the product that justifies the contract, more than the raw copy quality.

The second observation is about the sign-off flow. Work arrives where marketers already live — in my case Slack — with the predicted performance ranking attached, so approving a variant is a single click and the rationale travels with it. What I look for in this category is whether the tool removes steps or adds them; enterprise copy platforms usually add review steps, and this one noticeably did not in the channel I tested.

I also paid attention to the proof the vendor publishes, because that is what a buyer will be asked to defend: a claimed +23.3% average engagement lift, a 97% win rate against human-written controls, 30 million-plus messages sent weekly, and named customers including Tesco, Costco, Sephora, Best Buy, Nestlé, United Airlines, Currys and Accor. Those are Jacquard's own figures and I could not independently audit them; the win-rate methodology in particular depends on how the control groups are designed. Integrations listed on the site include Salesforce, Adobe, Braze, Iterable, Bloomreach, Optimizely and Epsilon.

How the learning works

I also spent time on how the platform talks about learning. The vendor's claim is that models learn from three layers at once: the world (the 80B-result corpus), your brand (approvals and voice), and your customers (campaign history and behaviour). The world layer is the vendor's own asset and you cannot verify it directly. The brand and customer layers are where a buyer gains the most influence: the more approved copy and outcome data you feed it in the first months, the more the prediction ranking reflects your house rather than the market average. Teams that onboard with a thin campaign history should expect the early predictions to lean generic.

Where it falls short

This is an enterprise messaging platform with an enterprise sales process. The $95,000-per-year floor means agencies, SMBs and most mid-market teams are simply out of the market, and I could not evaluate the product without a sales conversation — there is no sandbox to poke at. The scope, while broader than the Phrasee era, is still messaging: it does not produce campaign creative at large (imagery, video, landing pages), so it complements rather than replaces a creative or marketing automation stack. The Personalised Campaigns tier's headline price is nearly double the base package, and the Audience Optimisation add-on's audience-based pricing makes the total contract cost hard to bound before you talk to sales. Finally, any "average lift" figure across customers is an average: your baseline, category and test discipline will decide what you actually get.

On reputation: the company has been through a rebrand, a funding round and a CEO change inside roughly a year. That is not disqualifying — the product visibly shipped through it — but it is the kind of continuity question a procurement team will ask, and the honest answer is that Jacquard is a repositioned Phrasee with new money and new leadership.

One caveat on the agents framing itself. "Agentic" is doing a lot of work in the current marketing of this product, and the practical reality I could observe is more modest: three specialised checking and generation layers with an orchestration step on top. That is a sensible architecture and it works, but buyers should calibrate expectations to workflows, not to autonomous colleagues.

Verdict

If you run messaging at genuine enterprise volume — tens of millions of sends, regulated claims, multiple regions and languages — Jacquard earns its price and the brand/compliance agents are the reason. Rated 3.6 for the rubric mix that follows: core functionality is deep within its lane but the lane is messaging rather than full marketing production (weighted 40%), value is defensible only at scale given the $95k floor (20%), usability of the sign-off flow was good in my hands-on look (20%), the integration ecosystem and named customer base are credible (10%), and the reliability signals are mixed because of the 2024 rebrand and leadership turnover (10%). For everyone below enterprise scale, the correct rating is not "expensive but good" — it is "not for sale to you".

Sources

• https://www.jacquard.com/
• https://jacquard.com/pricing
• https://www.jacquard.com/resources/news/theres-a-new-name-in-brand-messaging
• https://capitald.com/news/phrasee-rebrands-as-jacquard-and-launches-its-personalised-campaigns-product/
• https://www.businesswire.com/news/home/20240515472852/en/Personalised-Content-at-Scale-Phrasee-Secures-Further-Funding-to-Accelerate-Product-Roadmap-and-Undertake-Rebrand-While-Announcing-New-Leadership-Appointments
• https://phrasee.co
• https://toolchase.com/tool/phrasee
• https://powermtaproviders.com/providers/phrasee.html
• https://staymodern.ai/solutions/phrasee/detailed
• https://platform.tracxn.com/a/d/company/5319f7b9e4b0f7e16604d87b/jacquard