As the potential around generative AI (GenAI) continues to build, business leaders are eager to harness this transformative technology. Successful GenAI projects begin with high aspirations, but completion requires careful planning and preparation that includes the computational support of cloud and lots of data — proprietary and third party.
GenAI models are trained on high-quality, curated and often proprietary data that is continuously refreshed. The key to meaningful, trustworthy data-driven insights includes thoughtful data operations, intelligent data management and an adaptive data platform.
According to HCLTech’s latest cloud research, Cloud Evolution: Mandate to Modernize, 98% of organizations are interested in GenAI that is trained on, or makes inferences based on, their proprietary data that can be verified and validated. Companies want to use their own data to train models, not just licensed data available to all.
To get their data management in order, there are four crucial steps that need to be taken:
1. Conduct a knowledge management session to identify all the relevant data sources and knowledge that will be required to train the GenAI model.
2. Assess the quality and accessibility of this data, using tools to ensure it is fit for purpose. This may require consolidating data from disparate sources into an adaptive data platform.
3. Determine if there are any ethical, regulatory or intellectual property concerns around using certain data sets. Establish a process for explainability and justifying the data being used.
4. Create a plan for continuously integrating new data sources and retraining GenAI models on a regular cadence. This ensures the model stays up-to-date and relevant.
Explainability and ethical concerns
The sources of proprietary data include patients, customers, partners and employees. Depending on the business there are always concerns about protecting the data, ensuring that what you collect will be used to benefit the related group. As businesses prepare to leverage the data, the challenge is explainability — can we explain how access to the data will benefit the providers of the data and can we ensure that we are adhering to ethical boundaries. “Do no harm” is an oath that can be borrowed from the medical profession as organizations prepare to collect and use data for GenAI solutions.
Additionally, the evolving regulatory environment, with regional differences, must be closely monitored. Establishing clear principles and processes for data selection, model training and ongoing monitoring is essential to mitigate liability risks and maintain trust as GenAI solutions are deployed.
Invest in platform engineering for cloud-native confidence
GenAI and AI use cases have supplanted traditional cloud use cases like burst capacity, test/dev environments and backup and disaster recovery.
“As organizations adopt generative AI, the need for skilled cloud-native developers is rising. To address this, many are investing in platform engineering, which automates complex cloud-native processes and enhances the management of applications. This strategic investment ensures more efficient development workflows and robust application scalability, reflecting deeper confidence in their technological advancements and future readiness,” said Pawan Vadapalli, Corporate Vice President and Global Head, Digital Business Services at HCLTech.
According to the research from HCLTech, 78% of organizations agree that cloud-native approaches to application development drive improvements. The cloud-native applications are scalable, infrastructure independent and portable with self-service deployment. Cloud-native application development accelerates response to the demand for GenAI solutions. Some organizations are experiencing a shortage of cloud-native developers. Consequently, 76% of the companies surveyed have invested in platform engineering. These platforms automate many of the cloud native processes.
“There is a correlation between the establishment of platform engineering and cloud-native confidence. We saw in the research that organizations extensively employing platform engineering reported a 22% (9 percentage point) increase in the proportion of cloud-native projects that they were managing,” said Vadapalli.
What about talent?
One challenge that can delay a GenAI project is the availability of professionals with AI experience and having capabilities to choose and train GenAI models like Google Cloud Gemini. For initial projects, IT should consider a partner with engineering and GenAI experience, as well as a knowledge of cloud-native application development.
As a service provider, HCLTech’s GenAI Labs offer an experiential environment that accelerates knowledge transfer and organizational confidence. We can close the skills gap while the organization’s team develops in-house talent.
Another advantage offered by HCLTech is our GenAI Industry Solutions. We are developing GenAI solutions to address various industry use cases. We select the model, address the data readiness of the business, train the model and orchestrate access to actionable insights.
How is HCLTech unique?
HCLTech differentiates itself from traditional consulting firms and other systems integrators based on its engineering heritage and hands-on experience in data management, GenAI and industry use cases.
Unlike firms that may rely on third-party resources to fill gaps, HCLTech’s own engineers have the technical expertise to tackle the complexities of data unification, model training and platform engineering. Our partnerships with cloud giants like Google, Microsoft and AWS, as well as AI leaders like NVIDIA, give us a distinct advantage, enabling us to stay ahead of the curve.
HCLTech also brings extensive domain knowledge from its decades of experience in business process outsourcing (BPO). This allows us to guide clients on starting their GenAI journeys with familiar, well documented use cases, such as order entry or inventory management. These types of applications are relatively risk free when it comes to data.
Starting a project based on a process you know without data that challenges regulatory guidelines enables you to start smart and finish faster. You can then build on that success and provide in-house developers with hands-on training.
A multicloud future
Looking ahead, there are other technology trends that are shaping how organizations adopt and mature their GenAI ambitions. According to HCLTech’s research, 87% of respondents are not only subscribing to cloud but have selected multiple cloud service providers. They site improved cloud security and access to emerging technologies that are only available on cloud as the top reasons drivers for multicloud. The survey also reported that hybrid cloud will persist.
As the data required to power custom GenAI solutions is distributed across on-premises and cloud infrastructure, the cloud infrastructure for these solutions will remain hybrid.
To realize GenAI ambitions, organizations must prioritize data readiness, ensure the computational performance required by GenAI solutions and align business requirements to the unique capabilities of cloud platforms. There is an overwhelming need for speed. No company wants to get left behind so partnering with an experienced systems integrator that understands both the technology and business processes is essential for turning GenAI aspirations into high impact solutions for the business.
