Tag: digital

AI and Automation: Graig Paglieri of Randstad Digital on How to Effectively Harness AI Technology in People Operations | In the News

Recently, Graig Paglieri, the CEO of Randstad Digital Americas was interviewed by Medium, where he talks about how they’re utilizing new technologies to make their jobs easier and provide greater strategic value.

In his interview, Graig mentioned a recent white paper with the Everest Group that estimates that the current global skill gap for AI/ML technologies is 25%-30% — and for cloud skills and cybersecurity, that skill gap stands at 20–25% and 30% percent, respectively. As a solution, enterprises should clearly outline what they hope to achieve with AI, from improving internal operations or enhancing customer engagement.

Read more in Medium.

India’s Wipro Likely to Skip Pay Hikes for Top Performers in Key Business Line – Memo | In the News

Wipro may skip giving hikes to “top performers with higher compensation” in its largest business line in the upcoming round of salary revisions in December, according to an internal memo seen by Reuters.

Such a move might affect employee morale and lead to more attrition, according to Peter Bendor-Samuel, CEO at Everest Group. “The additional attrition will help rebalance benches to meet demand, and will help bring cost back into balance with demand,” he added.

Read more in Reuters.

Humans at the Heart of Generative AI | In the News

Generative AI is becoming a key component of business operations and customer service interactions today. According to Salesforce research, three out of five workers (61%) either currently use or plan to use generative AI in their roles.

The public release of generative AI technology over the past year has improved the use of Chatbots dramatically in a short time. “Chatbots were around before, but generative AI has further increased their efficacy, as well as the quality of output,” notes Vishal Gupta, Vice President at Everest Group.

Read more in this MIT review.

Enterprise Generative AI Adoption: Risk Evaluation for Competitive Advantage | Blog

The adoption of generative AI technology poses four major types of threats to enterprises: data privacy and security, reliability and explainability, responsibility and ownership, and bias and ethics. By assessing current risk levels and implementing practices, tools, and systems to manage these challenges, enterprises can realize the most value from this transformative technology. Learn more about evaluating generative AI risk to gain an edge in this blog.  Learn more about our Generative AI Risk Assessment.

Generative Artificial Intelligence (AI) has captivated popular imagination like nothing else, promising a future filled with endless possibilities. For the first time, this technology can create art, synthesize human voices, and generate human-like responses to questions.

Open AI’s ChatGPT triggered the mainstream adoption of generative AI, racking up more than 100 million monthly active users within just two months of its launch. Today, more than 300 startups are developing various generative AI-related applications.

Enterprises globally have recognized generative AI’s emergence as a watershed moment and are scrambling to identify the best way to leverage its capabilities. Numerous use cases across industries and functions have already emerged and are being piloted.

Many technology providers have incorporated generative AI as an integral part of their solutions, and others are forging relevant partnerships to jump on the bandwagon.

However, while many organizations are excited about long-term generative AI adoption, few fully consider the potential risks. In this blog, we will delve deeper into the importance of generative AI risk assessment.

To realize maximum value from generative AI adoption, enterprises must undertake a structured incremental approach (as illustrated in Figure 1). This framework involves prioritizing use cases, assessing adoption risks, identifying suitable providers, adapting existing operating models, providing effective governance and change management, and reviewing performance against expectations.

Figure 1: Generative AI adoption framework
Figure 1: Generative AI adoption framework

Generative AI risks

Generative AI’s ease of usage has accelerated its adoption, highlighting both its value and its risks. Broadly, generative AI risks can be grouped into four categories: data privacy and security, reliability and explainability, responsibility and ownership, and bias and ethics (as shown below in Figure 2).

Figure 2: Generative AI risk categories
Figure 2: Generative AI risk categories

Let’s look at how these risks typically manifest and some examples:

Data privacy and security: Regulatory fallout from undisclosed data collection and retention is a key issue with generative AI models. This stems from the practice of developing AI models that can address a broad range of topics, rather than training data for a specific purpose. Further concerns include employees inadvertently sharing confidential enterprise data through user prompts or training data. In some cases, unfiltered prompts may allow employees access to data beyond their purview. From a cyber threat perspective, generative AI raises the risk of data breaches through malware, phishing, and identity theft

Samsung employees pasted confidential source code into ChatGPT to look for errors and optimize the data, inadvertently adding it to ChatGPT’s training data pool that can possibly be accessed by others.

Reliability and explainability: The quality and representativeness of training data greatly influence the accuracy of output produced by generative AI models. Deficiencies in the training data manifest as errors in generated content that may have serious legal ramifications beyond eroding customer trust. Furthermore, in the absence of required information, generative AI models may even fabricate information to answer a question. This leads to a false sense of expertise and can mislead the average user. Without a confidence score that estimates the likely accuracy of the generated content or some other equivalent mechanism, enterprises will need to develop and operationalize fact-checking of AI-generated content

During Microsoft’s Bing chat demo, the search engine was asked to analyze earnings reports from Gap and Lululemon and in comparing its answers to the actual reports, the chatbot missed some numbers and made some up. 

Responsibility and ownership: The legal ownership of a piece of content produced by generative AI raises complex questions. Does it belong to the enterprise that licensed the generative AI product or the company that owns the generative AI product? Moreover, do individuals or organizations whose content was used to train the AI model partially own any subsequent content produced by the AI? These legal quandaries currently lack clear answers. An evident problem is generative AI producing output that contains distinct and identifiable pieces of Intellectual Property (IP) owned by others. This can lead to potential legal fallout for the entity that deployed the generative AI model. Enterprises need to work with their legal teams to evolve their IP management amid widespread generative AI adoption

“Zarya of the Dawn” is a graphic novel written by Kris Kashtanova who used an AI based image generation software called Midjourney to create illustrations for the novel. After having initially given full copyright protection for the novel, the US Copyright Office later restricted the copyright to only the text and the arrangement of the illustrations and not the illustrations themselves. The justification provided was that copyright protection could only extend to human creators. 

Bias and ethics: An AI trained on biased data will propagate those biases, potentially leading to the generative AI producing discriminatory and stereotypical content. Failing to identify and preemptively remove such content through effective moderation can lead to severe reputational and legal ramifications for the enterprise and the generative AI provider.

Widespread generative AI adoption has the potential to ramp up carbon emissions from training and operating AI models. This can have significant implications for an enterprise’s Environmental, Social, and Governance (ESG) goals

In a study conducted by Bloomberg on Stable Diffusion (an AI-based text-to-image software), the rendering of more than 5,000 images for people with high- and low-paying jobs was full of racial and gender stereotypes. The results indicated men and individuals with lighter skin tones accounted for most high-paying roles.

How can enterprises assess their risk exposure to generative AI?

While the risks emanating from generative AI usage are notable, its benefits are too significant for enterprises to ignore. Consequently, enterprises that can leverage generative AI’s strengths while effectively mitigating its risks will outperform their peers. To effectively draw up a risk management plan for generative AI, enterprises need to first assess their current risk exposure to generative AI.

Everest Group has developed a multi-dimensional risk assessment framework (see Figure 3) to help enterprises take stock of their current risk profile for generative AI adoption. This framework is deployed through a tool that comprises 21 questions spanning the four risk categories mentioned above.

Figure 3: Everest Group’s generative AI risk assessment framework
Figure 3: Everest Group’s generative AI risk assessment framework

Responses provided by the enterprise across the four categories are weighted and aggregated to arrive at a risk score (see Figure 4).

Figure 4: Generative AI risk assessment outcomes
Figure 4: Generative AI risk assessment outcomes

Evaluating the risk exposure from generative AI is a necessary step to successfully implement and leverage generative AI to create value for customers. Incorporating appropriate risk management practices, tools, and mechanisms in the generative AI ecosystem can instill the confidence needed to take bigger bets, create differentiation, and fully harness this transformative technology.

Deploy our Generative AI Risk Assessment Tool. To discuss this tool and generative AI adoption strategies, please reach out to: [email protected], [email protected]; [email protected]; [email protected]; [email protected].

Check out our 2024 Key Issues webinar, Key Issues 2024: Creating Accelerated Value in a Dynamic World, to learn the major concerns, expectations, and trends for 2024 and hear recommendations on how to drive accelerated value from global services.

Festive High for Online Lending; Slowdown for IT Majors | In the News

With the festive season setting in, the digital lending industry is betting big on a quick rebound in business as consumers prepare to loosen their purse strings and merchants stock up to meet the additional demand. However, Bigwigs of the US$245-billion Indian IT industry may be staring at their slowest growth ever, data has shown.

“Some of the companies do risk posting their worst growth ever in 2024,” Peter Bender-Samuel, CEO of Everest Group, told ET.

Read more in The Economic Times.

11 Reasons Why Digital Transformations Fail, Explained by Pros | In the News

Digital transformation has come a long way from being a buzzword to becoming imperative for business success. However, digital transformation failures continue to plague many businesses, even as their organizations invest heavily in transformation efforts.

In a comment in Tech Target, Nitish Mittal, Partner at Everest Group, said that he sees many digital transformation initiatives struggle due to a lack of executive sponsorship. “Digital transformation initiatives, especially the major ones, need executive sponsorship, syndication, and backing,” he said.

Read more in Tech Target.

Exploring Emerging Generative AI Trends in Technology | Blog

Generative Artificial Intelligence’s rapid evolution holds the promise to transform enterprise operations and decision-making across many industries. Several emerging key generative AI (GAI) trends can profoundly impact automation, productivity, and human expertise, but harnessing GAI’s many opportunities will come with risks that will require enterprises to make complex choices and strategically adapt. Read this blog for valuable insights to prepare for this new frontier. 

Developing Generative AI Trends and Innovations

The trends to watch in the near and mid-term:

  • The move from general to specialized models – As generative AI moves into specific industries and domains, more examples of models fine-tuned for specific purposes are expected to emerge. For instance, models could be specifically trained for banking, insurance, or Human Resources domains, with the capability to speak the language of these narrower fields
  • Applications built on top of foundational GAI models – Apps built on top of large language models (LLMs) or conditioned LLMs to solve for specific needs will likely proliferate. Beyond ChatGPT, we already see early-stage web navigation concierges, code development assistants, and more. Initially, business-to-consumer (B2C) contexts will rise, but once the risks around GAI are solved, business-to-business (B2B) or business-to-employee (B2E) applications also will surge in activity
  • Lower costs – GAI is still relatively expensive but prices already have dropped significantly. As infrastructure, hosting, training, and inference become more efficient and economies of scale improve, we expect further cost reductions

What the generative AI trends mean for enterprises

  • Automation, productivity, and skills – Automation of tasks by GAI will boost employee productivity and also change the nature of expertise. This shift will require enterprises to rethink their talent agenda, workforce planning, learning and development (L&D) programs, and so on. Consider the example of an entry-level developer. With the benefits of GAI, the traditional “skill” of knowing a particular syntax for a specific language will become much less important. As a result, the bar of “valuable” human expertise will be raised. Enterprises need to account for these changes by rebuilding skill taxonomies and subsequently reassessing talent planning
  • Focus on enterprise data strategy – The true power of GAI comes into play once enterprises go beyond the low-hanging fruit of using it to generate generic outputs, like text, images, or other media. For instance, we could envision a world where GAI creates appropriate business or IT workflows, creates complex documents from scratch, or generates marketing collateral tailored to a company. Getting to these use cases will require seamless access to enterprise data, regardless of the approach (whether specialized models built from scratch, fine-tuning, or in-context learning). While GAI will unlock the power of this data, enterprises will need to surface it for use. The enterprise data journey is not new, but GAI will require a renewed focus and potentially more investments to advance it
  • Competition, disruption, and lowered barriers to entry – As GAI enables significant automation, organizations can do more with less. With lower costs, fundamentally new business models will become more feasible in multiple domains. Similar to how digital banks, built from the ground up, started nipping at the heels of established brick-and-mortar ones, this technology can potentially give birth to new contenders. One possible scenario to imagine is a new video game company creating complex video games relying heavily on GAI with a dash of human ingenuity. Similarly, GAI has the potential to disrupt stock media, customer service, entertainment, and other industries.

Enterprises may face difficult future choices, including making massive pivots, cannibalizing existing revenue streams, etc. While these decisions will naturally be difficult, enterprises must be willing to make hard calls to rapidly evolve and stave off existential threats further down the line.

However, there is no need to press the panic button yet. By investing in leadership education, keeping on top of developments, being open to innovations, and investing in home-grown and external GAI solutions, enterprises can position themselves well for when the time comes to make those hard choices

But before putting the horse before the cart, the many primary risks around GAI need to be addressed for broad-based enterprise adoption. These include regulatory concerns (including intellectual property), data and privacy, explainability (to some extent, at least), and others. Based on early trends, at least partial workarounds or mitigation mechanisms will be developed, in the short-term.

Everest Group provides insights and guidance on the risks, use cases, pricing, and implementation strategies to best position enterprises across industries for GAI adoption success. To learn more about Everest Group’s generative AI research or to discuss generative AI trends, reach out to Anil Vijayan.

Don’t miss our webinar, Key Issues 2024: Creating Accelerated Value in a Dynamic World, to hear our analysts discuss major concerns, expectations, and trends for 2024.

How can we engage?

Please let us know how we can help you on your journey.

Contact Us

"*" indicates required fields

Please review our Privacy Notice and check the box below to consent to the use of Personal Data that you provide.