September 2, 2020 By Dakshi Agrawal 3 min read

Language is constantly evolving – it shifts to meet the changing dynamics of our society. The business world is no different – it has its own vernacular that evolves in response to new innovations, changing consumer expectations and world events.

The language of business is documented in many enterprise forms: starting from simple text to more complicated formats like charts, tables, PDFs, and images – and each industry modifies the language of business to suit its needs. Without the right tools for interpretation, the meaning of language will be lost. Enter AI, in particular Natural Language Processing (NLP). NLP is the key to interpreting the trends and insights hidden within enterprise data.

AI that speaks the language of your business

At IBM, we speak the language of your business – and understand that each business has its nuanced terminology. That’s why IBM Research has put such a focus on developing and expanding IBM’s enterprise NLP capabilities: to help businesses unearth insights, answer questions and make more informed decisions that will allow them to solve whatever challenge is placed in front of them.

IBM Research creates a pipeline of NLP innovation that is continuously integrated into Watson products. This past March, we announced that we were taking some of the core NLP technologies powering IBM Research’s Project Debater – including advanced sentiment analysis (idiom understanding), summarization and topic clustering — and commercializing them within IBM’s NLP products, like Watson Discovery.

Today, we’re offering a first look at an NLP breakthrough that we plan to commercialize inside Watson products. This new capability will empower businesses to further understand and derive real value from their business data, so they can make more informed decisions, and provide customers and employees with more efficient insights.

The challenge businesses are facing: too much data

Businesses constantly commit to make or break operational decisions such as: pricing adjustments, product evolution, new marketing campaigns, and inventory optimization. When an organization prepares to make decisions, it is crucial that relevant information is readily available to be analyzed. Without a clear view of relevant points and considerations, businesses may be making uninformed decisions.

Envision this scenario: a business is gathering feedback directly from surveys, reviews and social media, but also wants to pull insights from call center logs and other documents such as field reports. It is challenging to sift through tens of thousands of documents and extract meaningful and actionable insights. The business wants to know: what are the top concerns? What emerging issues do we need to address?

That’s what the latest NLP innovation from the IBM Research team strives to achieve.

“Key Point Analysis” is the next generation of extractive summarization, evolved from extractive summarization capabilities first used with Project Debater. Key Point Analysis is a new approach for summarizing comments, opinions and statements from a set of documents.

How does it work?

  • Key Point Analysis examines thousands of documents and prepares a list of relevant points made in these documents by selecting, grading, and filtering high-quality passages.
  • Next, it removes key points that are too emotional in tone, are incoherent or include redundancies.
  • From there, it identifies how many sentences in the documents support various key points before filtering them further to ensure the results are diverse and cover the full span of information included in the documents.
  • Finally, Key Point Analysis delivers a ranked list of summarized key points, alongside a prevalence score and associated sentences that demonstrate the argument.

Organizations are presented with a concise, data-driven list of information that they can quickly act on.

For example, a financial services organization wants to improve the customer experience for credit card holders. Using Key Point Analysis, they could assess their customer complaint-related data (call logs, social media, incident reports, etc.) and easily identify top complaints. Key Point Analysis would offer them the following view:

  • Incorrect information on credit reports, including payment dates and amounts owed (17%)
  • Repeated calls and unwarranted contact (15%)
  • This account is fraudulent (7%)
  • My debts were paid in full (6%)

With this information, the financial services organization could decide which challenge to address first – in this case, incorrectly reported credit information – and quickly work to improve based on customer feedback.

Key Point Analysis empowers organizations to extract meaningful – and actionable – insight from the data – and evolve business decision making to ensure it is data-informed.

Interested in seeing key point analysis in action? You can participate in a new debate show, That’s Debatable, facilitated by natural language processing from Watson. Join the discussion by submitting a short argument. Get involved here:

Statements regarding IBM’s future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only.

Was this article helpful?

More from Artificial intelligence

Generative AI meets application modernization

2 min read - According to a survey of more than 400 top IT executives across industries in North America, three in four respondents say they still have disparate systems using traditional technologies and tools in their organizations. Furthermore, the survey finds that most executives report being in the planning or preliminary stages of modernization. Maintaining these traditional, legacy technologies and tools, often referred to as “technical debt,” for too long can have serious consequences, such as stalled development projects, cybersecurity exposures and operational…

Accelerating responsible AI adoption with a new Amazon Web Services (AWS) Generative AI Competency

3 min read - We’re at a watershed moment with generative AI. According to findings from the IBM Institute for Business Value, investment in generative AI is expected to grow nearly four times over the next two to three years. For enterprises that make the right investments in the technology it could deliver a strategic advantage that pays massive dividends. At IBM® we are committed to helping clients navigate this new reality and realize meaningful value from generative AI over the long term. For our…

How IBM and the Data & Trust Alliance are fostering greater transparency across the data ecosystem

2 min read - Strong data governance is foundational to robust artificial intelligence (AI) governance. Companies developing or deploying responsible AI must start with strong data governance to prepare for current or upcoming regulations and to create AI that is explainable, transparent and fair. Transparency about data is essential for any organization using data to drive decision-making or shape business strategies. It helps to build trust, accountability and credibility by making data and its governance processes accessible and understandable. However, this transparency can be…

IBM Newsletters

Get our newsletters and topic updates that deliver the latest thought leadership and insights on emerging trends.
Subscribe now More newsletters