Editorial
The Gist
- AI adoption. AI is set to revolutionize customer experience, boosting efficiency and savings.
- Data importance. The success of AI tools hinges on comprehensive, relevant training data.
- Enterprise ready. AI solutions must meet industry standards for security, governance and reliability.
We are living on the cusp of an AI revolution in customer experience management, as companies brace for a future where AI technology — such as machine learning, chatbots and lately generative AI — will supercharge contact centers and frontline teams and transform the way organizations manage their customer and employee experiences.
Integrations with new large language models (LLMs) and generative AI tools like ChatGPT will be irresistible for organizations that must do more with fewer resources and smaller teams. Generative AI can automate insight curation and summarization, or build a more intelligent, humanlike service bot, saving time and money.
But without a robust CX program foundation and a repository of useful omnichannel customer experience data to train the AI models, CX pros will find some of these tools do not have the firepower they need to deliver the kind of precision and accuracy their business requires to stay competitive and differentiated.
FOMO With a Side of Enterprise Scrutiny
Over the past few months, I’ve met with C-level executives and customer experience leaders at companies in nearly every industry vertical and they’ve all shared similar sentiments. When it comes to AI, they have a massive fear of missing out, and they are absolutely inundated with pitches for solutions and strategies claiming to provide the best path forward.
On the flip side, these leaders are trying to understand what it will take to make generative AI solutions “enterprise-ready,” meeting the myriad industry standards for security, governance, explainability and — most importantly — reliability.
ChatGPT may be the fastest-growing consumer app in history, but a bank, a government or any other regulated entity will have vastly different barriers to entry compared to your average user. They cannot risk proprietary, sensitive data being exposed or used to train a public LLM and becoming accessible to the public. These questions about data privacy, security, residency and compliance are the most significant obstacles to widespread enterprise adoption of generative AI.
Related Article: AI, Privacy & the Law: Unpacking the US Legal Framework
AI Is Only as Good as the Data Training It
The generative AI technologies dominating recent news cycles — such as ChatGPT and Bard — are trained on vast public data sources culled from the internet. By now, you’ve probably heard about the “hallucinations” that plague current generative AI models. This refers to situations where the model provides biased, false or misleading responses due to incomplete training data or a misunderstanding of the users’ prompt.
Forty-two percent of CX pros surveyed in a recent XM Institute study believe their organizations will invest significantly more in AI and machine learning solutions this year. An enterprise looking to deploy a generative AI solution to manage key business workstreams like customer service or brand health monitoring must ensure the output is accurate, predictable and repeatable.
To extract the most value out of LLMs, organizations will need to feed these models with industry-specific data and business context to develop solutions that deliver the precision and accuracy specific to their needs.
That means that organizations need to accelerate investment now to prepare data sets for LLMs to analyze that can power specialized, accurate and impactful generative AI solutions. Indeed, nearly half of the CX pros surveyed said investing in customer journey and text analytics are among their organizations’ highest priorities and that number will likely climb in the coming years.
Collecting and unifying experience and operational data will form a foundation to extract the most value out of LLMs, feeding models with industry-specific data and business context to develop solutions that deliver the precision and accuracy specific to their needs.
Related Article: AI-Augmented CX: How Customer Service Teams Can Learn and Adapt Faster Than Ever Before
Generative AI Solutions Built for the Enterprise
The difference between a purpose-built AI solution powered by data sources optimized for experience management and a basic LLM trained on publicly available data is akin to the difference between the expertise of a seasoned heart surgeon and a med-school graduate.
Take customer service. With the right training data, AI-powered solutions will do more to help agents answer generic customer questions with pre-programmed responses. It will help customer service teams understand every conversation on multiple human dimensions and quickly create personalized content, helping every frontline worker respond more effectively across service channels.
Marketers are exploring how they can use natural language prompts and AI-driven recommendations to create engaging content for different audience segments. Suppose an AI solution has the proper context and historical data points. In that case, it can go beyond generating a generic email draft and instead, generate personalized messages and offers to share with the right people at the right time, with pinpoint accuracy.
It is still early days in the generative AI revolution and there is time for enterprises to set themselves up for success. With the right data and customization, generative AI tools can be sharpened and honed to tackle even the most complex or specified business problems in nearly every industry.
Learn how you can join our contributor community.
