Accurate Image-to-Form Transcription

I have a stack of scanned images, each containing one to five fields that are primarily text—names, short codes, dates—along with the occasional numeric string. I will pass the files over in small batches, supply a brief field-by-field guide, and give you the direct link to my online form. Every character on the page must appear in the form exactly as it does in the image: correct capitalisation, spacing, and, most importantly, any special characters that show up. Extra spaces, added punctuation, or dropped zeros will cause the form’s validation to fail, so precision is vital. Once you have the layout memorised, I expect a steady rhythm: submit a batch, wait for my quick spot-check, then move on to the next. The job is finished when the final image is entered, the form throws no validation errors, and a random audit of records matches the source images perfectly. Please tell me the turnaround time you can consistently keep and highlight any previous work you have done converting scanned images directly into web forms.