proposeAI

Freelancers do not lose Upwork jobs because their skills are wrong. They lose them because every proposal reads like the last one. Writing a genuinely tailored pitch, one that reflects the client brief and the freelancer's specific experience, takes fifteen to twenty minutes per job. Multiply that across the dozens of jobs a serious freelancer applies to each week and the math stops working: either the proposals stay generic, or the freelancer runs out of hours for actual billable work.

proposeAI started from a simple observation: the freelancers winning the most work were not the most skilled, they were the ones who could write a sharp, specific pitch fastest. We set out to build a product that gives every freelancer that same speed advantage, without asking them to write from a blank page every time.

Every user builds a skill profile once: past work, specialties, rates, tone of voice. When a new job post comes in, the engine matches the relevant slice of that profile to what the client is asking for, instead of dumping the entire résumé into every pitch. The result reads like it was written for that one job, because effectively it was.

The generation engine parses the job post for scope, tone, and hidden requirements clients often bury in the second paragraph. It drafts a proposal that opens with the client's actual problem, not a generic greeting, and closes with a concrete next step. Users can regenerate in a different length or tone in a click if the first draft is not quite right.

A lightweight research pass looks at signals in the job post, and the client's posting history where available, to surface a detail worth mentioning: a past project, a stated deadline, a budget tier. That one specific reference is often what separates a proposal that gets read from one that gets skipped.

Finished proposals export with one click, copy straight to the clipboard or download as a document. Pricing runs on a credit system that refreshes monthly: five free proposals to start, then Starter and Pro tiers for freelancers pitching at real volume. No seat licenses, no annual lock-in.

proposeAI now supports more than 2,500 active freelancers and has generated over 120,000 proposals. Users report an 85 percent lift in reply rate over their previous, hand-written pitches. It remains one of the fastest-growing self-serve products in our portfolio, with adoption driven almost entirely by word of mouth inside freelancer communities.

Within months of launch, proposeAI passed 2,500 active freelancers and 120,000 generated proposals. Users report an 85 percent higher reply rate compared to their old proposals, the kind of lift that turns a slow month into a full pipeline. The product runs entirely self-serve, with no onboarding call required.

Every Devyst engagement follows a structured process: discovery, architecture, build, and handoff. This project was no different. We aligned on scope, reviewed existing systems, delivered iteratively, and handed off with documentation and runbooks.