Get the early bird discount when you register now for KMWorld 2017 in Washington DC

KM 3.0: KM AND AI

This article appears in the issue June 2017 [Volume 26, Issue 6]
Page 1 of 2 next >>


   Bookmark and Share

Artificial intelligence (AI) is increasingly entering consumer space via chatbots on e-commerce sites and news portals, as well as enterprise workflow through automation and machine learning tools. We may well be entering the phase of KM 3.0 (KM augmented with AI), but we will still build on KM 1.0 (traditional organizational content/community approaches) and KM 2.0 (enriched with social media and ecosystemwide co-creation).

Cognitive computing has been harnessed in the enterprise for data modeling and integration (see “The future is now: cognitive computing throughout the enterprise today” by Jelani Harper, KMWorld, May 2017). Recommendation systems and anti-fraud alerting mechanisms have also been augmented by AI (see “What’s True of AI in 2016” by Dan Adamson,.) Legal tools have been developed by Kira Systems, data extraction from sets of contracts; RAVN Systems, (graph database for visualization); and Workflows Autto, (preparing first drafts of investment agreements) (see article “AI takes hold in the legal profession” by Judith Lamont, KMWorld, February 2017).

“We may now be witnessing an inflection point in the rate of development and fielding of AI applications,” said Dan Roth, a professor at the University of Illinois at Urbana-Champaign, speaking at the recent AI Summit hosted in Bangalore by Kalaari Capital. Case studies were presented of AI in action at GE and Google.

Google’s AI journey began in 2012. It now uses neural networks to improve search ranking by RankBrain, recognize objects by Google Photos, provide translations by Google Translate, and provide smart replies by Gmail. Google has created an open source library called TensorFlow for machine learning, as well as cloud capabilities for data scientists. For users who are not data scientists, Google provides a machine-learning API.

Drones have used AI to recognize and count trucks at a construction site in Japan. Japan’s largest real-time car auction site AUCnet and French car insurance company ANZ France use AI for making predictions in patterns of usage.

Fleet analytics and engine diagnostics are top priorities for GE Software. The combination of AI and Internet of Things (IoT) can lead to a wide range of improvements such as asset optimization (for reliability), operational optimization (for capabilities) and business optimization (for market intelligence). The “digital twin” (identified by Gartner, as one of the Top Five tech trends of 2017) offers digital modeling of physical assets, which can be combined with machine learning to monitor engine performance and reduce downtimes. GE is working on “deep physics” models to predict performance of micro-turbines, thermal effects and flow models.

Such developments throw up a number of interesting questions for the KM community. How does AI affect existing KM practices? Who drives AI in the knowledge enterprise? What are some metrics to assess the disruptive impacts of AI? How are startups entering this space, and what can knowledge managers do to engage with them? Two groups of business practitioners in India gathered recently at TiE Bangalore to discuss the synergy between AI and KM: the Bangalore KM Community (a monthly meetup of CKOs) and The Indus Entrepreneurs (a community of startups and investors).

The panelists included Nikhil Nulkar, KM lead at Happiest Minds; Balaji Iyer, KM practice head at Grant Thornton’s Shared Services Center; and Ved Prakash, chief knowledge officer at Trianz. Here are some of my key takeaways from the wide-ranging discussion and debate that followed.

Impact areas and tools

Robotic process automation (RPA) has already impacted clerical and auditor tasks and is moving up the value chain. Recruiters are also using AI to go beyond resume filtering in the hiring process. AI will make it easier to find out who are really the subject matter experts (SMEs) across the length and breadth of the rapidly changing workforce.

Knowledge managers are looking for predictive search capabilities, which not only help employees locate documents, people or CoPs, but also make recommendations to them on whom else they should contact, what other documents/CoPs are relevant, or what new trends and developments are emerging in the field. By such contextualized suggestions, AI will accelerate the productivity promise of KM in shortening the time it takes to locate relevant assets; in fact, such information may be pushed to the user even before needing to search for it.

AI can also help with knowledge visualization and presentation, so that business intelligence is visible to a larger community of employees rather than to only those with “quant” backgrounds. Conversation on the enterprise knowledge portal is an activity ripe for exploitation by AI; discussions on Yammer and SalesForce.com have been mined for pattern analysis and predictive capabilities. A number of tech players such as Microsoft have developed chatbot frameworks which can be harnessed by knowledge managers.

Ownership and culture

In some companies, AI initiatives are launched by the CKO—in others, by the CTO. Not all companies have the capabilities to build AI tools and platforms themselves, hence the opportunity to team up with tech providers and startups in that space. Many startups in Bangalore attend the KM Community meetups to discuss opportunities for collaboration with CKOs. See ourstory.com/2015/04/tips-for-corporates-to-engage-startups and kmworld.com/Articles/Editorial/Features/Start-up-skills-and-knowledge-are-prized-by-large-enterprises-86786.aspx.

Page 1 of 2 next >>

Search KMWorld

Connect