20
Views
1
Comments
Creating a Chatbot in OutSystems 11 for Searching Documents – Feasibility and Effort

Hi everyone,

I’m exploring the possibility of building a chatbot in OutSystems 11 that allows users to ask questions and retrieves answers from a folder containing multiple files (PDFs, Word, Excel, etc.).

I’m looking for insights on a few points:

  1. Feasibility: Has anyone already implemented something similar in OutSystems 11?
  2. Implementation approach: How would you approach extracting text from documents and making it searchable within OutSystems? Any recommended patterns or tools?
  3. Effort estimation: Roughly, how much development effort (person-days/weeks) would you expect for a minimal viable version of this functionality?

Any shared experiences, tips, or guidance would be greatly appreciated.

Thanks in advance!


2026-01-12 12-58-24
Ramaaaaa

Hello Braham Ayad,

Yes, this is absolutely feasible in OutSystems 11, though it requires integrating external AI services. OutSystems works well as the UI and orchestration layer, while document processing and AI reasoning are handled via APIs.

For document text extraction, OutSystems does not provide native capabilities. You will need a document intelligence service such as Azure AI Document Intelligence or Google Document AI. These services extract text from PDFs, Word, and Excel files.

To make the content searchable and context-aware, the extracted text should be converted into embeddings and indexed for semantic search. This enables relevant document sections to be retrieved instead of relying on keyword search. (Optional but it would give you accurate  if you do this)

For question answering, integrate an LLM such as OpenAI (ChatGPT or GPT-4). The user question and the retrieved document content are sent together to the model, ensuring answers are generated strictly from your documents.

From an effort perspective, a minimal viable implementation typically takes 1.5 to 2.5 weeks(for me), covering document ingestion, API integration, semantic search, chatbot UI, and basic testing.

In summary, the solution is practical and scalable in OutSystems 11 by combining document intelligence APIs + LLM APIs, with OutSystems coordinating the flow and user experience.

Community GuidelinesBe kind and respectful, give credit to the original source of content, and search for duplicates before posting.