How we counted.
Round 1 followed a handful of private testers, and Round 2 traced Workday forms ourselves. Round 3 reads what happened when early-access users applied to real jobs on their own, with extension 1.0.3 and 1.0.4 in Chrome and Edge, in the 30 days to 7 November 2026.
- The source. Each time someone clicks fill, the extension records which fields it saw and how each one ended: filled, flagged for checking, or left for the person. On 7 November 2026 we ran one aggregate report over those records. It counts outcomes and prints no answers.
- The sample. 180+ applications and 4,500+ form fields. Per-application records begin on 15 October, so every figure except the field rates covers that part of the window.
- Anonymous by construction. The report prints only totals, and every figure combines at least five accounts. Nothing on this page describes a single person.
- What “filled” means. Jobsistant wrote a value into the field and read it back. That includes values it marked “Check this answer” in the panel.
Real use by early adopters, counted from the product’s own records. It is not a controlled study or a benchmark against other tools, and it records whether a value went in, not whether it was the answer the person would have chosen. That is why every form ends with your review. Round 1 and Round 2 used different builds and methods, so read the three rounds side by side rather than adding them together.
Your details went in almost every time: email 98%, phone 91%.
These are the ten fields with the highest fill rate. A rate is how often the field was filled when a form asked for it, across every form that asked.
| Field | Filled when asked |
|---|---|
| Full name | 100% |
| 98% | |
| Last name | 97% |
| First name | 96% |
| LinkedIn URL | 93% |
| Phone | 91% |
| Employer, in work history | 90% |
| Job title, in work history | 87% |
| Skills | 86% |
| City | 82% |
Some fields stay yours on purpose. Voluntary self-identification questions are never filled from your profile, and answers that depend on the job, such as salary or sponsorship, are filled for you to check rather than quietly.
Saved answers did more of the work: 1 in 20 fields on a first application, 1 in 5 by the 4th and 5th applications.
When you type an answer Jobsistant did not have, it saves it for next time. This table shows the share of each application’s fields that were filled from those saved answers, grouped by how many applications the person had made with Jobsistant.
| Application | From saved answers |
|---|---|
| 1st | 5% |
| 2nd | 9% |
| 3rd | 14% |
| 4th–5th | 20% |
Each row counts the applications at that point in people’s own histories, so it shows what reuse looked like at each stage across the group rather than one person’s path.
When a question matched a saved answer, it went in 73% of the time.
A saved answer is matched to a new question by its wording. When a match was found, Jobsistant filled the field from it 73% of the time; a looser match is filled and marked for you to check. Across the whole sample, saved answers supplied 43% of everything Jobsistant filled.
97% of typed values held when Jobsistant read them back.
Job forms are built with many different widgets, and some quietly undo a value typed into them. So after writing each value, Jobsistant reads the field back, and flags any field where the value did not take so you know to look at it.
How to read these numbers.
- Field rates are per field. They describe the fields named, not whole forms. Forms also ask questions only you can answer, and those are left for you.
- Filled is not the same as checked. The records show that a value went in and stuck. Whether it was right is what your review is for.
- Small and early. Early adopters applying to the jobs they chose, not a random sample of job seekers or forms.
Disclosure: Jobsistant publishes these field notes and has a commercial interest in autofill that works. The figures are aggregates of our own product records, not normalized, peer-reviewed or compared with any other product. The earlier rounds cover how Workday forms behave and how long application forms run.