Connect with us

Hi, what are you looking for?


Neural net model will be the next frontier of competitive advantage

Manoj Agrawal interviews Gartner Vice President Van Baker

Manoj Agrawal: How many magic quadrants does Gartner publish for AI, ML & DL? Which was the last and which is the next?

Van Baker: On AI – None. The market is too immature for AI, ML & DL. We have magic quadrants on analytics. We try and avoid DL as terminology and focus on DL network. We won’t do magic quadrants on AI overall, but we have guides on virtual customer assistance, hosted AI services, and conversational platforms. The earliest we can have a magic quadrant on AI is at least 12-18 months away. Lots of AI companies have good awareness and revenue…but not profitability or cash flow.

What differences do you see between AI offered by big tech vs well-funded start-ups vs less funded start-ups?

From a ML perspective and broad application perspective, you need a lot of resources – CPU, memory & data – especially for unsupervised learning. That limits it to large companies like Google, IBM, etc. Or else someone who has the ability to tap resources on cloud. Even a small AI requires big resources for the design, but not for production. For production, the algorithm could run on a phone.

Early adopts will face the challenge of having adopted a weaker AI technology in some of their cases. How can they figure out when their choice is no longer the best and what challenges will they face in shifting from one AI tech to another?

Some providers in the market tightly couple chabots with their own NLP (natural language processing) engine. We think there is risk associated with that. We are starting to see the emergence of middleware in the API space. Chatbot providers like Liveperson have a lot of integration capability and bot building capability, but you can bring bots built using other frameworks and bring them into their environment. So flexible frameworks are emerging. When I talk to Amazon, Google, etc, about how often they completely rebuild the AI engine and create a new algo, the answer is every 6 months. So, it is important to have flexibility.

The same AI tools are available to everyone – so will AI change the nature of competition? How?

It will. In this era of digital biz, where the pace of biz is changing very rapidly and the application of AI will give insights that will give a brief competitive advantage. The competitive advantage is not going to be sustainable and there will quickly be level a playing field. Data has no value, so differentiation will come from the building the neural net model, which is a very challenging task. Neural net model will be the next frontier of competitive advantage – and that is a challenge for data scientists and experts to build unsupervised / supervised / reinforced models. It is a mix of intuition and science.

How can one distinguish between intelligence and wisdom – so as to not lose wisdom to AI – eg wisdom about coding or software architecture or usage insights or vulnerabilities or human psychology?

Machines don’t think. AI is algorithms. They are digesting a lot of data and identifying relations. All this is machine learning. ML can handle much larger data sets compared to humans. But what is missing is common sense. ML is actually pattern recognition.

Are VCs using AI to analyse investment proposals?

It could be applied to a portion of the analysis. VCs look not only at proposed product and its features and capabilities, but also for leadership qualities and people skills.

Which sector of financial sector is leading the effective deployment of AI? Investment banking.

[email protected]

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

PR Newswire

Copyright © Glocal Infomart Pvt Ltd. All rights reserved. Usage of content from website is subject to Terms and Conditions.