Oct 12, 2017 | Claire Cunningham
How to Drive Product Innovation with Natural Language Generation
In an effort to ensure their products are at the forefront of the technological landscape, businesses will turn to a variety of platforms, plugins, and extensions that they believe will help them stand out and be perceived as innovative among competitors. Oftentimes, the most lucrative and disruptive technologies come with an understood (and daunting) level of risk. While product investments can oftentimes come with these large-scale risks, integrating premier artificial intelligence (AI) technologies such as Advanced Natural Language Generation (Advanced NLG) has proven to generate even larger-scale rewards.
A subfield of AI, NLG produces human-sounding language from data at scale. The integration of Advanced NLG into technology products and platforms can be a major differentiator by broadening the adoption of your product, enhancing smart data discovery capabilities, and driving the creation of new products. When applied successfully, the additional value your employees and customers can gain from this integration are enormous.
Consider the following questions to make an educated decision on if and how to integrate NLG to drive product innovation:
Are your customers sitting on treasure chests of unused data?
Global revenue for big data and business intelligence analytics platforms is estimated to grow from $122 billion in 2015 to a staggering $187 billion by 2019. In this world where companies are spending billions of dollars collecting data on their customers, potential customers, competitors, and their own performance, it is crucial to find a way to make that data valuable. NLG can efficiently add value to a product where this “big data” is not being utilized by generating consistent narratives that highlight actionable insights that your organization and customers find most interesting.
For example, one company we work with is a large-scale accounting firm that files international tax returns for global organizations and logs hundreds of thousands of individual taxpayer data points. Originally, this data was stored, unused, and forgotten after tax season. However, by integrating NLG into their new product, each and every taxpayer now receives a highly customized narrative with a detailed analysis on why they have a balance due or a refund. They also receive advice on how to improve their finances for the next year with little to no lift from their individual tax practitioner.
Is there a problem worth solving with this solution?
Hot new AI technologies sound extremely appealing in terms of the “perceived” benefits. Make sure you are asking the following questions to understand the value Advanced NLG can bring to your product:
- Do you need to communicate data analysis in a consumable way to various audiences?
- Are you required to perform high frequency, high volume reporting?
- Are you spending too much time analyzing, interpreting, and communicating data insights?
- Do you need to communicate custom information at scale?
- Are you looking to broaden adoption of your BI or data visualization products?
For example, one of our customers is an auditing team that performs on-demand fluctuation analysis detailing the drivers of variation in financial statement line items to highlight strengths and weaknesses for a given company. By incorporating NLG into this workflow, the team is enabled to send different versions of the same report to members of the C-suite, senior-level managers, and business analysts based on each person’s required level of information. The team is also freed from the time-consuming task of manually analyzing and compiling the relevant insights, and the logic for this type of work can be applied across as many financial statements as necessary.
Is there already some level of automation in the workflow?
NLG is great at maximizing the reach of your solution when there is already some level of automation in the workflow. Making sure the necessary data is sent to an NLG platform and distributed to the relevant audiences is arguably one of the most important elements to consider when thinking about a possible integration. Ideally, your solution will be incorporated into an existing digitized process, meaning the data required is already being collected and the distribution channel for the narrative is in place.
For example, if the data used to drive your product is already automatically compiled and being shared to other visualization or BI tools, that same data can be sent to an NLG platform to provide the corresponding text and insights.
There are a number of other factors to consider when evaluating a possible partnership between NLG and your products, but the above considerations outline the pillars of success for your future endeavor.
Read our research brief to learn more about how NLG fits into the Artificial Intelligence eco-system: