Cloudera: Business Leaders Fret Over Generative AI Issues on Data Privacy
New research from data company Cloudera shows that although 53% of US organizations already use generative AI technology and an additional 36% are exploring AI use for potential implementation next year, 84% of decision-makers worry about data privacy, security, and compliance. Also, 95% believe that full control of data during AI model training is key to trusting the AI outputs.
“Organizations are apprehensive about the potential exposure of training models using publicly available data and/or receiving erroneous responses from AI models that have NOT been trained with relevant enterprise context,” said Abhas Ricky, chief strategy officer at Cloudera, the Santa Clara, California-based provider of a platform for enterprise data management and analytics.
Related Article: Cloudera Announces New Observability Solution
The main benefits of generative AI, according to the research, are enhanced customer communication through the use of chatbots or other tools (55%), support for product development and concept development (both tied at 44%), and support for data analysis (34%). The findings can be found in the Cloudera study, 2023 Evolving Trends: Data, Analytics & AI.
Verint: US Stores Will Face Staffing Shortages and Other Issues for the Holidays
More than half of US retailers believe that the effective staffing of stores will be the greatest challenge this year for the upcoming holiday season, new research shows from Melville, New York-based analytics and customer engagement firm Verint.
At least 86% will face one challenge related to the hiring and staffing of customer-facing roles, 50% will have difficulty finding qualified people to work in their stores, and 30% will struggle to fill positions in the contact center and back office. Moreover, 34% will not have enough human resources (HR) staff to handle the vacancies that need filling.
In such an environment, 55% of business leaders consider chatbots to be a key part of their customer engagement strategy, up 10 percentage points from last year. But beyond the labor scarcity issue, retailing success will depend on connecting customer data across all channels and putting customer feedback into practice, the research findings indicate.
Although retailers recognize the value of collecting customer feedback to understand buyer behavior, only a slim majority feel that insights are being used effectively, and many struggle to share customer data across channels to support omnichannel engagement. However, those able to leverage insights effectively are also able to improve CX in-store (73%) and on digital channels (75%), the study reveals.
Talkdesk: AI is Key to CX Success, But CX Teams Have Work to Do
AI-powered CX presents an effective yet relatively untapped opportunity for financial organizations facing pressure to retain and grow customers during economic uncertainty, the new Talkdesk 2024 CX in Banking Report reveals.
Throughout the industry, CX has become an increasingly strategic priority in the past 12 months for 98% of banks and credit unions. It is regarded as a leading driver of customer loyalty within the financial services space (83%). Yet although AI can dramatically improve customer service, CX teams must first allay internal concerns about the technology.
Related Article: eGain and Talkdesk Partnership Elevates Agent and Customer Experience
Per the report, the biggest barrier to integrating AI into an organization is resistance to change (66%), followed by the insufficient availability of talent (63%). Larger organizations are also concerned about lack of quality data (74%), indicates the report, which comes from San Francisco-based Talkdesk, the provider of an enterprise contact center platform.
Even so, more than half of respondents (56%) expect their company to increase investment in CX technology by 10% or more during the next three years. The top planned AI investments include analyzing contact center data for actionable discoveries and insights (83%), optimizing chatbot or guided conversations (73%), and virtual customer assistants for self-service (66%).
Stanford University: More Transparency Is Needed for AI Foundation Models
A report from Stanford University researchers measuring the transparency of AI foundation models urged companies such as OpenAI and Google to disclose more information on the data, human labor, and other factors used to train the models.
Companies are developing foundation models that serve as AI systems trained on massive datasets to perform a multitude of tasks, because of surging interest in generative AI following the launch of ChatGPT by OpenAI. Because these models are increasingly relied on for decision-making and automation, it is crucial to understand their biases and limitations, the authors of the report said.
“It is clear over the last three years that transparency is on the decline while capability is going through the roof,” says Stanford professor Percy Liang, a researcher behind the Foundation Model Transparency Index. “We view this as highly problematic because we’ve seen in other areas like social media that when transparency goes down, bad things can happen as a result.”
The index graded 10 popular models on 100 different transparency indicators, such as training data and how much compute was used. All models scored “unimpressively:” The most transparent model, Llama 2 from Meta, received a score of 53 out of 100, whereas Titan from Amazon ranked the lowest, scoring 11 out of 100. GPT-4 from OpenAI obtained a score of 47 out of 100.
The authors of the index hope that the report encourages more foundation model transparency. The index is a project of the Center for Research on Foundation Models (CRFM), a new interdisciplinary initiative emerging from the Stanford Institute for Human-Centered Artificial Intelligence (HAI) to study foundation models.
Kore.ai: Intelligent Virtual Assistants Help Resolve Service Frustrations
A new CX benchmark report from enterprise AI platform provider Kore.ai reveals a growing acceptance of AI-powered intelligent virtual assistants (IVAs) that meet customer service quality expectations. US customers in banking, health, retail, travel, and telecom/cable/media increasingly rely on IVAs to get prompt and relevant information while engaging in conversation, without having to repeat themselves, the report said.
Customers value the efficiency of self-service and desire a quality experience, but they do not want to sacrifice quality for convenience―all while demanding accuracy, speed, and expertise. Thanks to the vast improvements in technologies powered by conversational AI, generative AI, and large language models (LLMs), IVAs can now meet these needs. Even so, users appreciate an easy handoff to a live agent when needed, according to the report.
Among the findings: US consumers still favor a live agent (77%) over an IVA (70%), but the preference for IVAs is growing. Moreover, 66% prefer getting an order status in 30 seconds from an IVA, versus waiting for three minutes for a live agent. And 79% of consumers loathe repeating their situation or problem, citing this as their biggest customer service frustration.
Author Information
Alex is responsible for writing about trends and changes that are impacting the customer experience market. He had served as Principal Editor at Village Intelligence, a Los Angeles-based consultancy on technology impacting healthcare and healthcare-related industries. Alex was also Associate Director for Content Management at Omdia and Informa Tech, where he produced white papers, executive summaries, market insights, blogs, and other key content assets. His areas of coverage spanned the sectors grouped under the technology vertical, including semiconductors, smart technologies, enterprise & IT, media, displays, mobile, power, healthcare, China research, industrial and IoT, automotive, and transformative technologies.
At IHS Markit, he was Managing Editor of the company’s flagship IHS Quarterly, covering aerospace & defense, economics & country risk, chemicals, oil & gas, and other IHS verticals. He was Principal Editor of analyst output at iSuppli Corp. and Managing Editor of Market Watch, a fortnightly newsletter highlighting significant analyst report findings for pitching to the media. He started his career in writing as an Editor-Reporter for The Associated Press.
