NashTech Insights

Winning business with AI – Being Strategic

Ram Indukuri
Ram Indukuri
Table of Contents

This is the first in the series of enterprise AI blogs by the Advisory Group within NashTech.

Large corporations are getting serious about AI, thanks to the recent excitement created by ChatGPT. Speculations abound about what AI will do to various industries. Some are predicting AI will eliminate the need for several classes of jobs. This triggered CEOs and Technology executives, to spend time understanding the impact of AI on their company. At this point, most companies have some investments made in AI, solving specific problems. Progressive CEOs are asking the technology executives to propose how the company could benefit from AI. However, very few companies have been strategic about it.

What is Strategic AI Capability?

Being strategic is to believe that AI is here to stay and invest in the AI capabilities of the company. Ultimately, the capability should make the company repeatedly build products powered by AI that help gain market share or reduce costs dramatically.

Unfortunately, most companies are looking at AI as a tool. Ie., to solve a specific problem and approach it tactically than strategically. Some customers, particularly those who are excited about ChatGPT, reach out to us to answer the question, ‘What can AI do for us ?”. AI can do a lot of things. The best people who know what AI can do for you is your employees. But they should be empowered with the knowledge of AI.

The following picture depicts the 3 components of enterprise AI capability.

AI Culture

AI Culture is no different from innovation. Be adventurous, stick your neck out, learn every day, be curious, take calculated risks, forgive failures, and demand action. This is no easy task, especially, AI has a bad reputation as a a technology that is going to eliminate jobs.

Efficiency (Company/Internal Focused) Back in the 80s and ’90s, the Japanese implemented quality circles, where in workers join together and eliminate waste systematically. That led to a major manufacturing efficiency increase across the board and laid the foundation for Japanese car manufacturers to win the market share. The same will repeat for AI and the reason will again be the participation at the grassroots level in implementing AI.

Products or Services (Customer Focused) While the above para explains how the culture can impact ‘internal‘ efficiency, which needs buy-in from all employees, building AI into your products, requires your company to adopt to product management processes, with capable product managers. This requires, transformation in how you build products or services in your company, and as a business executive, you need to drive your teams to think ‘products’.

An $8B SaaS customer in the HR space, embarked on a journey to AI, by funding a team that is tasked to implement an AI-powered recruitment process. While the team delivered great results and learned the art of production-grade AI with the help of NashTech, the same could not be percolated to all the product teams and the AI was more or less centralized.

AI Skills

Lack of AI Skills is the most difficult obstacle in implementing AI. As the world is racing to adopt AI, people with AI backgrounds are in great demand. A structured approach to educating your staff with the help of a Consulting company like NashTech would go a long way in getting your company’s AIQ higher.

A 27B USD retail customer, implemented AI by building a great data science team, however, was less focused on the MLOps and other essential skills. The projects suffered engineering failures and had to engage NashTech to bring in engineering and program management skills.

AI Skill Levels

AI Assets

The Key AI Assets are Data and Tools. Data is the oxygen for AI. Hence a systematic way to obtain, curate and make it available is fundamental to any enterprise AI program. Companies need to take a look at their ownership of data sources. They need to work backward to ensure all legal and infrastructural hurdles are addressed. A strong data foundation is a necessary prerequisite for a strong AI program for the enterprise.

Data Program: Modern Neural Networks require a significant amount of good-quality training data. Unfortunately, Most companies do not have a good data program that provides the data for AI projects. It is understandable, the urge to get an AI use case out the door and not have to deal with the wider challenge of instituting a data foundation at the company. In those situations, we advise to create a simple infrastructure. Infrastructure is limited to the domain that is in question, on a cloud of your choice. Major cloud providers, have made it very simple to have a data foundation that is a few hundred dollars to begin with and scale as you use more. They also bring the best practices of data along with them. This indeed solves the tool selection side of AI assets as well. The following diagram illustrates the data mesh architecture, that enables AI programs.

Computational Infrastructure: The computational resources required to implement AI are expensive and currently scarce. About 25,000 A100 GPU devices are used in training GPT4. Companies are acquiring Nvidia GPUs at a faster rate. As a company, you can rely on cloud providers, but they are going to be prohibitively expensive. You may need to look at buying hardware for your local data centers. At the moment, the capability to build infrastructure and invest in computational power is a critical factor in winning AI race.

An ambitious $1.8B recruiting firm, embarked on an AI journey but suffered slow progress due to a lack of data foundation. The AI team had to deal with identifying the source systems, convince the business owners of applications to share data, curate and manage the life cycle of data before they could start using it for training the AI models. The customer engaged NashTech to build the data platform and had to put off AI program temporarily and eventually succeded.

NashTech AI Methodology

NashTech has combined about a decade of experience implementing AI into a structured methodology in going about implementing AI. This methodology includes all the elements discussed above.


The benefits of approaching AI strategically are multiple.

  • The guard rails laid by the data foundation will ensure that your data is safe and secure.
  • The foundation enables, significantly reducing the implementation cost of each AI initiative. It enables teams to see results (Failed or Successful) quickly, thus building more momentum.
  • Executive buy-in ensures longer-term productivity improvements and increased market share, beyond the day-to-day hype about AI
  • As a by-product, your organizational innovation improves.
  • You can draw parallels with Digital Transformation. Companies that invested in building digital products or modernizing customer experien eventually benefited from those processes. These capabilities seep into their traditional IT and business.

For more information about NashTech AI Methodology, contact Ram Indukuri (

Ram Indukuri

Ram Indukuri

Head of AI & Advisory Consultant

Leave a Comment

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

Suggested Article

%d bloggers like this: