Generative AI

Incorporate the power of Large Language Models (LLMs) like PaLM 2 and Vertex AI into your applications. We have developed applications that use and fine-tune LLMs for:

  • Document search and updating,

  • Secure Chatbots

  • Complex pattern detection

  • Student evaluation

  • Real-time context analysis for voice

Chatbots are a popular example of AI used in business, however this is just a small example of what’s possible. AI can be used to deliver a better customer experience, optimise processes and provide businesses with deep and accurate insights.

Computer Vision

We build solutions that analyse images and video in real-time that deliver insights related to:

  • Object detection (Cloud Vision API)

  • Face detection (Cloud Vision API)

  • Video summarisation and context extraction (Speech-to-text Transcription Extraction)

  • Transcript extraction (Speech-to-text Transcription Extraction)

  • Word detection (Cloud Vision API)

Natural Language Processing (Text Analysis)

We analyse text to perform:

  • Sentiment analysis (NLP API)

  • Topic and intent classification (NLP API)

  • Translation (Cloud Translation API)

  • Mood classification (NLP API)

Personality Profiling

We are able to use behavioural traits to predict personality profiles based on OCEAN and Myers-Briggs, DISC and others.

Network Analysis

We apply Neo4j to analyse social media networks and other abstract graphs for a number of insights. These include:

  • Influencer analysis

  • Information campaign detection

  • Prediction of success of information campaigns

Audio Analysis

We utilize Speech-to-text to analyse audio in real-time to:

  • Classify sounds

  • Extract specific speaker transcripts from noise environments

  • Measure and remove noise

 
 
 

A Few Success Stories

Automated Document Analysis

The Problem to Solve

Our clients would like to perform automated document analysis and bulk analysis of other data sources using advanced artificial intelligence.

Our Solution

Documents were ingested into GCP for processing. Then all text components were analysed using the Natural Language API for sentiment, entities, sentiment towards entities. A number of custom machine learning models were trained and deployed using Vertex AI to classify text and posts according to certain criteria. Results were stored in BigQuery and made available as Looker dashboards.

The outcome

A highly scalable and specific solution was created that enabled highly-detailed insights to be gleaned, in an automated manner, from complex documents.

Semantic Search and AI evaluations

The problem to Solve

Our clients would like to automate the evaluation of learner submission that involved the analysis of images and complex text content and to allow learners to use natural language to search video and textual materials.

our solution

The solution involved the submission of materials using an API to the artificial intelligence evaluation capability. Diagrams are automatically analyzed using machine learning models developed by the AddAxis team. Large Language Models were fine tuned and used for the analysis of text and images. Vertex AI language capabilities played a major role in the development of the solution.

the outcome

Learner submissions are automatically evaluated using machine learning and are able to search videos and documents using semantic search to find the most relevant context.

 
 

Generative AI for Business