TECH OFFER

Chatbot Training Platform

KEY INFORMATION

TECHNOLOGY CATEGORY:
Infocomm - Human-Computer Interaction
Infocomm - Big Data, Data Analytics, Data Mining & Data Visualisation
Show more >
TECHNOLOGY READINESS LEVEL (TRL):
LOCATION:
Singapore
ID NUMBER:
TO174371

TECHNOLOGY OVERVIEW

Bots are supposed to understand what humans want, guiding them to their desired outcome. However open-text chatbots tend to do poorly because they are unable to detect intents (what the customers are actually asking) accurately.

Our technology puts intelligence into our clients' chat or voice bots, to know what their end-users want.  Our proprietary rules-based models predict user intentions by (1) automatically analysing human-bot conversations and (2) using this to train bots faster, better and in a more cost-efficient manner. 

Using our platform takes 20x less human intervention. Consequently, bots become smarter much faster than if clients relied on manual training, typically reducing errors by 68% in as little as 3 weeks. This brings the current  average chatbot accuracy rates from 30-40% to 90+% within several training iterations. The platform also provides tools to help clients identify customer trends and their emotions towards the bot.

We are looking for corporate clients and bot developers who want to differentiate themselves in the marketplace by offering chatbots that actually meet their customer metrics using our proprietary training platform.

TECHNOLOGY FEATURES & SPECIFICATIONS

We offer training of chatbots using our proprietary NLP/NLU on an automated basis, resulting in accuracies of 90% or more. By automating this process, we can increase the training instances and develop company specific NLUs faster and more accurately than via manual processes.

We have a proprietary NLU/NLP that supplements mainstream NLPs to better identify missed intents and false positives in client chatbots. The client uploads the chatlog and knowledge base via API to our platform and the platform analyses the chatlog to auto-detect (1) false positives, (2) false negatives, and (3) recommended answers to a high degree of accuracy. The plaform is also able to help (1) group unclassified questions to form new intents, and (2) identify cases where semantically alike utterances lead to multiple answers, allowing the knowledge base to be de-conflicted (usually around 20% in a mature chatbot knowledge base).

We currently are able to support English, Singlish, Thai, Bahasa Indonesia, Simplified Chinese and Vietnamese. NLUs for most structured languages can be developed within 30-60 days.

POTENTIAL APPLICATIONS

  • AI Chatbot training - increases answer accuracy from traditional 30-40% to 90+%.
  • AI Voicebot training - increases answer accuracy from traditional 20-30% to 90+%.

Other potential applications are in areas where there is a need to understand human input and to match it with an appropriate answer.

Market Trends & Opportunities

The world needs better bots. More people are shopping from home during late hours. Contact centres are being flooded with requests but can't staff enough agents to service those questions. Our bot-trainer - develops your existing bots and turns them into delightful sales and customer service experiences (4-star rating in a scale of 1-5) https://www.a-star.edu.sg/News-and-Events/a-star-innovate/innovates/innovate/a-star-spinoff-intnt.ai-trains-chatbots

Benefits

Bot owners are able to train their chatbots regularly and at far lower costs than doing so manually. By ensuring that their bot is accurate and can engage their end-clients well, customers are able to realize the gains promised by bots - lower customer service overheads (a reduction of 72% of calls deflected to a human support staff), a superior sales engine (increase of up to 400% in sales leads) and actionable customer trends and insights.

RELATED TECH OFFERS
Accelerated Materials Innovation Platform for Sustainability
Real Time, All-day, Stress Monitoring System Using Data Science
AI-Based Medical Imaging Assistant For Breast Cancer Screening
Improving Explainable Artificial Intelligence For Degraded Images
Enhancing Construction Safety and Productivity with Video Analytics
Smart Manufacturing Video Analytics via CCTV
Smart Buildings Video Analytics via CCTV
Universal Autonomous Navigation Software for Robots
Intelligent Video Analytics to Enhance Construction Worksite Safety
Smart Cities Video Analytics via CCTV