BCG: Five Rules for Fixing AI in Business

In business, using artificial intelligence (AI) and machine learning is sexy.  But is it practical?

A recent Boston Consulting Group* (BCG) article says, “For AI to have real value in an organization, both its potential and its limitations must be understood by the business side of the company as well [as by technologists], so that expectations and outcomes can be realistically established.”

According to BCG, “Using machine learning in a business setting has by and large been discouraging.  A 2020 study conducted by BCG GAMMA, the BCG Henderson Institute, and MIT Sloan Management Review found that only one out of ten companies has enjoyed significant benefits from AI.  Try as they may, a company’s data scientists cannot often repeat the successes of machine-learning competitions in day-to-day operations in which decision making must be forward looking and reliable.”

To rectify this issue, BCG proposes five rules companies should adhere to as they look to implement AI and machine learning to advance real business processes.  While BCG intends its rules to guide building AI applications, it turns out they also apply to assessing the value of AI-enabled applications, such as Northern Light SinglePoint™.

BCG’s five rules are:

  • The AI application must match the desired business-related outcomes – BCG notes that often companies do not devote “the time needed upfront to fully understand the input data required by the specific business problem.” In the case of a market and competitive intelligence research portal like SinglePoint, that data typically includes both internally and externally generated research reports, news, and industry and government databases, all of which must be ingested, tagged, and indexed consistently and completely.
  • Use external data to amplify the business impact – BCG warns companies against adopting a “not invented here” attitude regarding data for building AI applications and training sets. Similarly, an AI-enabled research portal like SinglePoint should be populated with all relevant content from licensed third-party sources to ensure that decision makes throughout the enterprise can access insights to inform their decisions.
  • Conquer complexity by breaking down the AI model into its smallest parts – BCG advises companies to “put an AI model’s agility to work by designing simpler and more interpretable sub-models targeted at the business logic of relatively circumscribed problems.” Similarly, Northern Light has created thousands of meaning-loaded entities so that its text analytics can operate at a very granular level and distill meaning and insights from long, complex research documents, enabling researchers to work far more efficiently than would otherwise be possible.
  • Machine learning should help in making concrete business decisions – BCG suggests “performance [of AI] should be measured against the incremental value that the company gains from decisions driven by the system.” That’s certainly true for a knowledge management system like SinglePoint, in which the “deliverable” to users is relevant information responsive to a specific research inquiry.  The incremental value of such a system is faster time-to-insight, and better “concrete” business decisions.
  • Avoid machine learning outcomes that seem accurate but may not prove useful – This is equally true from the perspective of users of a knowledge management system like SinglePoint – it is not sufficient that a search result is “relevant” to a user’s inquiry, it must be “meaningful” in a way that advances the user’s understanding and insight. That is the user’s definition of “useful.”

# # #

* BCG is one of dozens of authoritative sources contained in Northern Light’s Thought Leaders Content Collection, available to SinglePoint subscribers directly from Northern Light, and to individuals and companies through Amazon’s AWS Marketplace and the Amazon Data Exchange.

Back To Top