Internal involvement in AI projects cannot be outsourced

Abhimanyu CEO and Co-Founder

In a post-AI world, the case for AI has been made already. As organizations make their “journey towards AI”, the term AI implementation is being discussed threadbare everywhere, from presentations to documents on hiring strategy. However, that doesn’t imply everyone is doing it right.

At a recent event, Arjun Maheswaran (Co-founder & Chief Scientist, Agara Labs) shared a few insights on what companies need and where they usually go off-track. In this freewheeling discussion, Arjun was joined by veterans from varied industries such as consulting, manufacturing, and urban design. We have presented a few snippets from what Arjun spoke during the discussion.

Arjun addressing other panelists at the ET AI-IoT Summit. Image Courtesy: ET Edge

Why do AI implementations fail?
“Although the CXO buys your product, the end-user needs to be your champion. Without the buy-in from the end-user, it is hard for the product or POC to be successful. At Agara Labs, we usually involve the end-user starting from the product development. By doing so, they know what they are getting into, how to use the product, and be a champion for it.”

AI still needs to solve a business problem
“We are not implementing technology for the sake of technology. We have seen companies fail to tie it to a business problem. For example, lack of awareness regarding what metrics need to be tracked, or failing to define the problem down to the last detail contribute to a less-than-satisfactory AI implementation.”

Don’t underestimate internal involvement
“Organizations tend to underestimate the involvement needed from within the company. This involvement cannot be outsourced to a third party or consultant because the employees are usually the domain experts. To illustrate, there is a lot of feature engineering required during product creation and employee involvement is essential.”

Is AI a core competence of the organization?
“This a critical question that needs to be answered for acquiring the necessary AI skill-set. If the answer is no, then the rule of thumb is to go with a third-party. However, one still needs an internal team that understands the business and works with the third-party.”

“If it is going to be a core competence, then the simplest way is through an acquisition. One should acquire a team in the right space and let them understand your business problem. Afterward, you can build a team around them, and do more acquisitions, if need be”

Abhimanyu’s take on the impact of AI in enterprises was featured in Economic Times (Page 19, 10th May 2019)

This panel discussion was a part of the ET AI-IoT Summit held in Mumbai on March 15. The other panelists in this discussion were :

Moderator: Anchal Arora, Director, Digital Consulting, KPMG

Sumit D Chowdhury, (Founder & CEO, Gaia Smart Cities)

Samit Dutta, (Head, Supply Chain, IT, Digital Transformation & Operational Excellence at Piramal Glass Ltd.)

Siddhartha S, (CEO, Intain Technologies)

Related Articles
May 31, 2019 4 Minutes
Agara Labs closes Pre-Series A funding

Agara Labs, a startup applying Deep Learning to speech and text to automate customer support, has come out of stealth and added to its Pre-Series A round

September 25, 2020 12 Minutes
Is a voicebot just a chatbot with voice processing capabilities?

This article highlights some of the challenges and considerations in designing and building a voicebot. Do you think Voicebot = ASR + Chatbot + TTS?

February 13, 2020 4 Minutes
The Newest Dimension in Customer Service

Has the seemingly far-fetched reality where we talk to AI-powered voicebots to get things done, just started getting real. Read more about this new dimension in customer service here.

Team Agara