How Lines of Business Will Prove the Value of GenAI

Businesses are now operating in a type of digital transformation trifecta: 1. The popularity of generative AI is rapidly driving data and applications to the cloud where large language models (LLMs) reside; 2. Emerging GenAI tools and services are fueling more SaaS growth; and 3. Lines of business (LOB) exploring how GenAI can enhance their productivity to develop competitive LLM applications for their customers. In this transformation cycle, it is too early to prove ROI for most use cases due to the lack of benchmark data.

Rapid Experimentation With GenAI

Never in the tech innovation cycle has another technology come close to hype and rapid adoption. Cloud transformation and SaaS application delivery has been marching along for some time. Over the past year, the landscape has changed dramatically with the introduction of GenAI, which is exponentially increasing cloud transformation. Businesses wanting to leverage the power of GenAI are looking for new ways to drive revenue through product development, nurture customer engagement with new product marketing tools or support internal operations teams.

Lines of business are at the forefront of experimenting with how GenAI can add value and will lead the way in accessing distributed GenAI services delivered via SaaS. Our team has also been experimenting with GenAI services for various use cases, most notably search and summarization capabilities for internal operations teams (DevOps, engineers, support and others).

Experiments Are Underway

One way to look at GenAI is to realize that all activities are still in beta. Refinements and new models change every week. About 90% of GenAI usage can be considered experimental, with about 10% moving into production. While early days, LOB teams are ready to get started with GenAI, seeing the demonstrated productivity in using it to write code, support development, search internal knowledge bases and aid in research.

Rapid experimentation means that data security and controls are not in place while teams implement these GenAI services. This pattern was seen in cloud adoption where security was late, and security teams had to scramble to put controls in place. For GenAI services, it will be beneficial to bring security teams along, since there is sensitive data being shared to new services in the cloud.

Extracting Value From GenAI

Integrating GenAI into a business must be filtered through the same considerations that apply to any development or services project, with the caveat that it is an emerging technology, and some traditional metrics may not apply. For example, a typical SaaS cost model will look different with GenAI, as off-the-shelf AI tools are not yet in any measurable deployment. One of the most popular use cases is to augment support teams with GenAI tools. The ROI here is to measure the deflection rate of support tickets, the volume of support over time and response time metrics. All of these metrics are being measured today, so one could argue that support teams have an effective ROI metric to measure the before and after with GenAI tools.

GenAI will increase in productivity and value over time. Here are key practices for LOBs to consider as they embark on their GenAI integration:

• Determine use case and services: Small-to-medium businesses with leaner staff need to be highly targeted when allocating time to GenAI. Starting with a use case is the right first step. The second step is to determine which services will be used. Whether it is code generation or data analytics services, starting with a carefully thought through development plan will prevent wasted time and resources.
• Understand the cost vectors: The cost of GenAI is not a neat linear equation, as its use is still largely experimental. However, a LOB team leader can look at the cost units involved, such as number of tokens, ingestion cost, number of queries and storage cost, to get a picture of the overall budget impact. In addition, there is a fixed subscription cost per user and how that scales as the service goes into production. It is also a good idea to factor in DevOps and other resources to maintain the service as with any production project.
• Gain visibility into GenAI usage: Businesses with limited IT staff will need the support of an easy-to-use, real-time monitoring and reporting solution to assess how GenAI is being deployed in their organization. Since experimentation is such a large part of GenAI at this point, it is important to be careful not to impose controls on usage too early and prevent what could become valuable development. Existing data controls, such as browser isolation techniques, can also provide visibility and monitoring for any anomalous activity.
• Secure GenAI data for privacy protection: To enhance the security of GenAI data and safeguard privacy, businesses must prioritize security, especially considering the alarming statistics. According to "The State of Data Security" report by our team, 41% of organizations across various industries have experienced a security breach in the past year. This underscores the critical need for robust security measures, especially for businesses utilizing LLMs, to avoid compliance violations and protect against potential data privacy breaches. Guardrails are becoming an imperative, as governments and businesses are concerned about LLMs using data sets that could violate personal privacy. Visibility monitoring into which GenAI tools are running in a business will not only identify how the tools are being used, but also help protect against data privacy infringement.
• Managing GenAI with a hybrid remote workforce: The hybrid workforce presents its own set of challenges, as many remote workers leverage GenAI. Using a remote browser isolation solution can help ensure that requests to access data comply with access and policy controls. Businesses must also work to democratize security by encouraging every person, remote or on-site, using GenAI tools be highly aware of queries and analytics that could access data subject to privacy protections. If privacy data is used to train a foundational GenAI model, the data breach could be perpetuated through various analytics projects, compounding the violation.

Over time, a business will be able to identify predictive usage patterns and develop restrictions based on those patterns. Now and in the future, visibility is key to avoiding compliance and privacy issues that are costly to any business.

GenAI and SaaS Lead the Way

Businesses typically talk about ROI when debating whether an innovative technology or application is worth pursuing. GenAI is just getting started, and it will continue to show new ways to deliver value. As AI moves farther along and benchmark performance data is collected, it will be possible to build demonstrable ROI models. For now, it is an exciting time of experimentation as new productivity enhancements will surface, fueled by the power of GenAI.

Learn how you can join our contributor community.