Unlock Market Insights with Precision
Online Market Research powered by RAG (Retrieval Augmented Generation)
In cooperation with Gysho we are proud to present the next generation Market Research Assistant, powered by:
- Retrieval-Augmented Generation
- A choice of GPT models: 3.5 and 4
- 20 research types and areas
- Comprehensive research reports
- Chat: interactive assistant
- Fully customisable for your product or organisation
Go to Gysho MarqtAI for a demo
Intelligent Chat Assistant: Streamline Your Marketing Tasks. Work with MARQTAI assistant to put the research data into practice.
- Create strategic plans
- Perform keyword analyses
- Create campaign ideas
- Generate content for; instagram, Facebook, LinkedIn etc
- Validate decisions
Accelerate with Marketing Insights: with MARQTAI you no longer need to rely on gutfeel or perform long studies to make decisions. Validate your decisions instantly.
AIWISE - Systems Design Engineering Project Requirements AI
The tool takes any large requirement Excel sheet, and will produce the same document with recommendations, scores and reasoning
- Pre-processing: language entity extraction to solve O(n) problem
- Vector DB for high performance preprocessing
- Prescribed syntax and standards fine-tuning
- Language agnostic: input document language determines output language, as supported by LLM
- Improves speed of QA and quality of requirement documents
- Optimises decision making process
TLDR; Analysing and converting project requirements according to SMART and INCOSE standards for Systems Design Engineering.
AIWISE employs Large Language Models to facilitate quality improvement of large requirement documents. The POC phase has just been finished and accepted, and new requirements are being developed for the next exciting phase: implementation!
Broad application of AI models. Talk to your data. It's simple!
TLDR; Try out the new demo! Upload a document, and have a conversation with your data!
This sample application demonstrates the use of context, while chatting with the most current and powerful Large Language AI Models. Additionally it is possible to chat with a local data set, leveraging the power of AI models on any type of data, in any language.
At WinkIT we believe in the power of AI to do tasks that are humanly impossible. The following applications have proven to be very successful with our approach:
- Question and Answering on your datasets, such as knowledge bases, PDF documents, Excel...
- Large data set manipulation, data pruning, grammar, spelling, and language style modification
- Customer support bots; using your organization's knowledge base
- Custom pipelines for automatic document mutation
- And much more!
Whether your organization needs Chat Bots, Knowledge Base optimization or translation, Custom Workflow optimization, WinkIT can provide a solution that will blow your mind. Curious? Don't hesitate to share your idea's and questions via the Contact page.
Tender Portfolio Management Platform
Bid Book Tenders, the ultimate in Tender management. state-of-the-art system is designed to help businesses and organizations of all sizes succeed in the highly competitive tender market.
The responsive and user-friendly interface currently supports:
- 6 years of data, going back to 2018
- Lightning fast search, with advanced filters
- Free three week trial for new users
- Ability to follow tenders, with email notifications
- Automate searches, with email notifications
- Tender prognosis - finally know what's going on
- Account management
The platform is still under heavy development. Written in the incredible Meteor Framework (Meteor, React, Mongo, JavaScript) it is the latest state-of-the-art Custom Web Application to come out of the WinkIT factory.
One of the main challenges for this project was performance. Working with big data-sets can be challenging when speed and accuracy is required. A lot goes on here, data pruning, data massaging, and storage is optimized to get the best performance possible.
The back-end features a continuous import of newly released Tender data, keeping users in the loop, managing user account expiration, and notifications for all users.
CMU Dict a custom linguistics word list generator
The CMU Dict tool is another custom web application that generates word lists that facilitates teachers in helping children with dyslexia; a learning difficulty that primarily affects the skills involved in accurate and fluent word reading and spelling.
The tool implements the Carnegie Mellon University Pronouncing Dictionary. An open-source machine-readable pronunciation dictionary for North American English that contains over 134,000 words and their pronunciations.
More information on the dictionary can be found here: The CMU Pronouncing Dictionary
The main challenge here was implementing the often inconsistent mapping of syllables and phonemes of the North American language pronunciation. This resulted in a rather large mapping table, but with excellent results!
The tool is coded (blast from the past!) in Perl, CGI with MySQL as the database back-end.
Custom scripts where written in Shell script to facilitate importing the CMU database into a usable and custom Relational Database schema.
The next step in this awesome teaching tool is to implement an AI Large Language Model to create short stories, using only the words supplied by on of the saved word lists.
This was an interesting project, as I had to get familiar with certain linguistic science principles. But mostly it gave me a good feeling, to be able to work with bright minds to create a learning tool that will benefit children with a learning difficulty.