Artificial Intelligence
Jun 21, 2019

AI: realize the value sooner with pretrained accelerators

Enterprise investment in artificial intelligence (AI) is ramping up. Genpact's AI 360 research revealed that within large companies, a quarter of senior executives say they plan to fundamentally reimagine their businesses with AI by the end of 2021.

Unfortunately, despite these intentions, the number of enterprises realizing maximum ROI from their AI projects is still low. In this blog, I'll outline the challenges organizations face on their AI journeys, explain how pretrained accelerators can help, and share some recommendations on how to lay foundations for success.

Breaking down barriers

With so many organizations investing in AI, you would expect to hear plenty of success stories. However, research from McKinsey shows a number of barriers are delaying or derailing AI deployments. These include:

  • A lack of talent with appropriate skill sets for AI work
  • A lack of clear strategy for AI
  • A lack of available data and limited usefulness of data
  • A lack of senior leadership commitment and ownership
  • Functional silos constraining end-to-end AI solutions

Despite this, businesses aren't giving up. The potential benefits of AI are so compelling that enterprises – as well as vendor and service provider communities – are innovating to break down these barriers.

For example, collaboration is essential. Any single enterprise will struggle to collate required data and combine it with industry and business process domain knowledge to effectively train an AI system to improve day-to-day operations.

But what if you could leverage pretrained AI algorithms rather than having to start from scratch? Today, what once seemed like fantasy is now a reality.

Enter pretrained AI accelerators

Pretrained AI accelerators allow enterprises to leverage the unique combination of domain knowledge, labeled data, and AI engines relevant to your business processes. As a result, you can realize the value of your AI investment sooner.

Forrester has already noted the benefits: "Accessing and preparing training data is one of the hurdles for AI adoption. Pretrained vertical AI solutions obviate the need for training because a vendor has already trained the model for a specific use case." Forrester goes on to encourage enterprises to invest in AI accelerators due to high potential business value, low cost to implement, and the fact that “the benefits in process optimization and consistency are almost immediate."

HFS Research also supports this view. "Business leaders can leverage the "AI in a box" approach to quickly build disruptive AI use cases for first-mover advantage, disrupting competitors at their own game. The key is to bring all relevant components required to bring AI solutions for your business together “in a box" so that you can quickly build, train and deploy the use cases, rather than reinvent the wheel."

About the author

Dan Glessner

Dan Glessner

Vice President, Digital

Follow Dan Glessner on LinkedIn