No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

For Gen AI to Succeed, Data Has to Come First

AI_2CND889_71922.jpg

AI technology
Image: Pitinan Piyavatin - Alamy Stock Photo

CX technologists have been trying to leverage enterprise data comprehensively since the dark days before any of us had heard of ChatGPT -- which is to say, pre-2022. Vendors in the customer experience and adjacent spaces have been pushing customer data platforms, with the vision that the more of your customers’ data that surfaces to your contact center agents at just the right time in the interaction, the more effectively you could serve, satisfy, and upsell that customer. Before any of this happens, however, that customer data needs to be cleaned and integrated to make it accessible to the contact center platform.

Generative AI takes the concern over data to a whole new level. Now the focus is not just customer data for CX systems — it’s all data that could be relevant to whatever use case you’re trying to serve with Gen AI. And if enterprises found it challenging to unify and integrate their customer data in the pre-Gen AI days, they’re finding the Gen AI-era data challenge even more vexing.

Solar Winds has just come out with a survey finding concerns over data quality to be a significant blocker for Gen AI adoption. The management vendor reported that just 38% of survey respondents “are very trusting of the data quality and training used in AI technologies, and [respondents] rank data quality as a major barrier to AI adoption, second only to security and privacy risks.”

Given the importance of data to the success of AI projects, we’re devoting an extended 90-minute “Deep Dive” session to the topic at our new Enterprise Connect AI event Oct. 1 – 2 in Santa Clara, CA. The abstract for this session presents the challenge succinctly:

Artificial Intelligence (AI) and Machine Learning (ML) projects can provide significant return on investment when they are applied to narrow but difficult business problems and supported by adequate amounts of relevant, quality data. Many such projects start off with high hopes but get derailed due to fundamental problems with source data which were not known to the organization before embarking on the work. Without some careful analysis prior to beginning work, these defects can crop up as blockers surprisingly late in the AI/ML development process.

The session will be presented by Joshua Powers, technical director for AI/ML at software development firm Dev Technology Group, and when we were planning the session, Powers emphasized the point in the above paragraph: The enterprise has to tackle the data challenge at the beginning of the process, or risk losing valuable time, if not failing completely, in a situation where the problems aren’t discovered until too late. Powers breaks the data challenge down into four elements, all of which have to be addressed at the front end of any AI project:

  • Knowledge Representation
  • Data Availability
  • Data Quality
  • Data Ethics

One of our goals for the October EC AI event is to, “Help make AI a reality for your enterprise,” and clearly, data readiness is a prerequisite here. I hope you can join us in Santa Clara this October for the chance to get practical insights from leading enterprises and industry experts, who can help guide you through the range of obstacles that organizations are facing as they look to prove in and scale Generative AI.