AI-powered paper survey and assessment digitization for research, education, healthcare, and enterprise teams.
Upload scanned or marked response PDFs from any survey template. Lens reads checkboxes, ratings, and handwriting, in whatever language your respondents used, against your survey's own schema, and hands you back data that's ready for analysis.
No new survey platform. No manual data entry. No response data used to train AI models. Lens reads the survey you already built. Learn more →
Typing up hundreds of scanned forms by hand is slow and error-prone. Generic OCR tools don't know your survey's questions or options, so they guess instead of matching. Every new batch means starting over.
Staff manually retype every checkbox and handwritten answer from scanned forms. Mistakes creep in with every hour of data entry. Every question type, matrix grids, rank orders, open text, needs its own workaround. There's no way to reuse the work across batches of the same survey.
Lens reads each scanned or marked response against your survey's own schema. Once a survey's layout is mapped, every batch of that survey reuses it. Low-confidence answers are flagged automatically instead of silently guessed. Fix one question or one response without redoing the rest.
The layout only needs mapping once per survey. Every batch after that goes straight to extraction.
Upload your Pirai AI base-format schema, or a raw questionnaire document and Lens derives one. This becomes the answer key every response is checked against.
Upload one blank or sample copy of the survey. Lens reads every page and proposes which questions live where. Review and confirm before processing real responses.
Drop in scanned or marked PDFs in bulk. Lens reads checkboxes, ratings, and handwriting, verified against the schema, not guessed.
Low-confidence or blank answers are flagged for review. Edit any answer directly, reprocess a single question, or reprocess a whole response, then export to CSV or Excel.
Convert the survey, map the layout from a blank PDF, upload a filled response, and export the result, start to finish.
Not just checkboxes. Lens reads the full range of question types a real survey uses, each matched against its own known answer key.
Multiple choice with one selection per respondent.
Checkbox questions where more than one option can be marked.
Single-selection questions presented as a dropdown in your platform.
Statement-by-scale grids, matched against every row and column.
Handwritten short answers and essay responses, transcribed verbatim.
Multi-field entries like version numbers or device details.
Items ranked by the respondent, read in the order marked.
Points or percentages allocated across a fixed list of items.
Lens covers the full digitization workflow, from reading the layout once to flagging exactly what needs a second look.
No fixed layout to design around. Lens reads the survey you already built, in the format you already print, and maps it to your own schema.
Reads handwritten short answers and essay responses, not just checkboxes, transcribed verbatim against your survey's own answer key.
Reads marks and handwriting in the language your respondents actually used. Not limited to English-language surveys.
Spot a wrong or blank answer? Select exactly that question and reprocess it, without redoing the rest of the response.
Every extracted answer is directly editable in the review screen. You're never limited to AI reprocessing, a human always has the final say before export.
Every answer is checked against your survey's own questions and options, so nothing gets invented or mismatched.
Lens reads every response against your survey's exact wording, so two similarly worded questions are never mixed up, even when an answer spans more than one page.
Low-confidence and blank answers are surfaced automatically, so review time goes where it's actually needed.
Export as values or labels, split multiple-answer questions into separate columns, headers matched to your platform's convention.
Pay for what you process. No separate platform fee.
Every scanned or marked page counts toward your total, whatever question types or how many questions it contains.
Email Us to Get StartedDigitize paper-based fieldwork, intercept surveys, and panel studies without re-keying a single response.
Course evaluations, paper-based exams and quizzes, and classroom assessments, scored and digitized automatically, no re-keying answer sheets by hand.
Patient satisfaction forms and paper-based intake surveys, digitized without a new platform to learn.
Employee surveys, compliance audits, and customer feedback forms collected on paper across locations.
Pirai AI Lens is Pirai AI's paper survey digitization tool. It reads scanned or marked paper survey response PDFs and extracts every answer, checkboxes, ratings, matrix grids, rank orders, and handwritten text, matched against your survey's own schema.
Lens uses AI vision models to look at each scanned page and report what's visible, checked against your survey's own known options and items rather than guessed freely. Answers Lens isn't confident about are flagged for human review instead of being silently recorded.
No. Lens works from the same schema Pirai AI already derives for 19+ survey platforms, so you bring the survey you already built or already ran, and Lens digitizes the paper responses against it.
Low-confidence and blank answers are flagged automatically in the review screen, so you can check exactly the answers that need a second look instead of reviewing every response from scratch.
Yes. Lens can reprocess a single flagged question in isolation, or reprocess a whole response, without disturbing anything else already extracted correctly.
Yes. Every extracted answer is directly editable in the review screen. You don't have to rely on AI reprocessing, a human always has the final say before anything is exported.
No. Response images are processed only to extract your survey data. Nothing you upload is used to train AI models. Learn more →
Single answer, multiple answer, dropdown, matrix, open text, form fields, rank order, and constant sum, eight question types in total.
Lens accepts scanned or marked response PDFs, and exports results as CSV or Excel, formatted with values or labels and headers matched to your survey platform's convention.
Lens reads handwriting and marks in the language your respondents actually used, it isn't limited to English-language surveys.
Lens is priced at $9.99 per 1,000 pages processed. Email us to get started →
Pirai AI Lens is built for teams still collecting responses on paper, across market research, education and assessments, healthcare, and enterprise. Send us a note and we'll walk you through how it fits your workflow.
