Product

Quescall

Answer questionnaires automatically — AI-assisted from your own answer history. No more hours of copy-paste on supplier, compliance and security questionnaires.

What Quescall does

Supplier self-assessments, security questionnaires, compliance audits, ESG reporting, RFP responses — many companies keep digging up and rewording the same answers for similar questionnaires. Quescall ends that: incoming questionnaires are ingested, every question is matched against your own answer history automatically — and matching answers are proposed.

  • Upload: PDF, Word or Excel — Quescall extracts the questions in a structured way.
  • Semantic search via vector embeddings (PostgreSQL + pgvector): finds topically matching past answers, not just exact text matches.
  • AI fallback: if no history fits, an LLM answers — your choice of Claude (Anthropic), OpenAI, Mistral or a local model (Ollama / LM Studio).
  • Review workflow: you check, complete or reject — Quescall never submits without your approval; every answer has a confidence indicator.
  • Learning: accepted answers flow back into the answer history and improve future suggestions.
  • Output: the completed questionnaire as a PDF (via Gotenberg) — download or e-mail delivery.

Quescall runs as a web application on your own server (on-premise). Answer history and embeddings stay in your own PostgreSQL database. With local AI models the system works entirely without any external cloud connection — also usable in regulated environments.

How Quescall is structured

Intake & extraction

PDF · DOCX · Excel

  • Upload of incoming questionnaires in common formats
  • Structured extraction of questions — including from PDF forms
  • Manage several questionnaires in parallel

Matching & AI

pgvector + multi-provider LLM

  • Vector search in the answer history via pgvector
  • AI fallback: Claude, OpenAI, Mistral or local (Ollama / LM Studio)
  • Confidence indicator per answer suggestion

Output & learning loop

PDF generator + AnswerHistory

  • PDF generation of the completed questionnaire (Gotenberg)
  • Delivery by e-mail or direct download
  • Accepted answers go back into AnswerHistory — the system keeps learning

Quescall runs as a .NET 8 application with PostgreSQL 16 (pgvector) in Docker. Hosting either with us, in your own cloud or fully on-premise — on request entirely without external model APIs.

Live instance at quescall.csbg.de — get in touch for questions about trial installation, on-premise deployment or custom adaptations.