Artificial IntelligenceInterviews

AI is Essential for Data Analytics and Data Science


Xiaodong Zhou, the CTO at Presight, says Generative AI have shown solid evidence and great potential in terms of content understanding

What have we achieved so far in terms of use case scenarios of Gen AI?
We are working aggressively on Gen AI-related POCs and Projects for short-term and long-term use cases, we consider some areas are low-hanging fruit, enterprise document search and Q&A, Chatbot and Digital Human for different companies and requirements. We have also developed enhanced NLP capabilities that with far better ability and accuracy for information retrieval and conversion from unstructured data into structured data. We use large imagery models to index and provide semantic search capabilities to large image datasets. We use LLM to improve the traditional specialized OCR model into the LLM-enhanced OCR process with faster implementation and better accuracy.

Why according to you should companies leverage generative AI?
Generative AI have shown solid evidence and great potential in terms of content understanding, this is critical for our data analytics and data science business, instead of a generic understanding of column A, B, and C and connection X and Y between datasets, now we can infer lots of information from the column name and values, be able to understand importance and relationships among data automatically and be able to map natural language based business questions into specific data queries and model building.

What are the challenges companies are facing in terms of adopting and using Gen AI and how can they be overcome?
Skillsets, not many people are involved in Gen AI development, so sourcing talent with solid skills or developing internal technical know-how is a challenge. But we see fast adoption of Gen AI and more and more people are gaining good Gen AI expertise every day.

Gen AI capabilities, it is not magic that Gen AI will be able to solve all problems, it has its strengths and weaknesses, we are also learning from baseline use cases with success stories and advanced use cases, sometime it is less known that can be solved by Gen AI. We just need to balance the efforts in terms of resource allocation to deliver something solid and useful but also risk-taking investment in areas less known.

Some issues like the Hallucination problem from Gen AI models are particularly important in enterprise Gen AI challenges we are learning various ways to overcome this technical challenge.

Are companies aware of regional and global policies surrounding the use of Gen AI? 
I believe there are enough concerns and warnings raised by experts, the government and large companies on Gen AI risks, from legal to privacy and copyright violations etc. I am sure with more and more adoption of Gen AI in enterprise use cases, along the way, a clearer picture of technology solutions, policies and regulations will be available for companies to adopt. We are always closely monitoring these particular areas of development.

How can companies use their resources on using Gen AI to create a competitive advantage?
I think this may be different for different companies, I personally will focus on one, how Generative AI can help our traditional business do better, faster and cheaper, and two, how Generative AI can bring new capabilities that differentiate us, it can be previously tried but failed practices, or previously thought impossible to achieve.

What factors do companies need to consider before adopting Gen AI such as having a centralised data strategy?
The short and long-term impact of Gen AI on business. Eg some low-hanging fruit implementation for process improvement, and better customer services, to how Gen AI has the potential to impact the company business model.

  1. Acquire skillsets or collaborate with experienced partners to invest in Gen AI solutions
  2. Potential ethical and legal impact
  3. Understand potential Gen AI limitations like hallucination and bias
  4. Potential Cost and ROI, eg cost of expensive GPU resources with performance and scalability limitations
  5. Data availability, model selection etc are also early technical questions that should be analyzed, but stay flexible as new models and solutions are coming out rapidly

How can companies experiment with Gen AI to predict the future of strategic workforce planning?
We believe Gen AI will have a profound impact on many businesses, from software to IT, from financial to legal, what we have experienced is to Start from concrete problems with subject matter experts support, and quickly build and validate with Gen AI capabilities. Instead of directly applying GAI on open-ended and broad questions, such as How can I use GAI in the financial domain? So it gives executives a clearer picture what can be the impact of adopting Gen AI. This will in turn help businesses to realize new opportunities, expand capabilities and also workforce planning, including new skillsets acquisition and training of existing workforces.

Prarthana Mary

Cybersecurity Defences Employing AI Can Combat Threats with Greater Speeds

Previous article

Bybit Achieves ‘AA’ Rating in CCData’s Crypto Exchange Benchmark Report

Next article

You may also like


Comments are closed.