There’s no getting away from it. The buzz around generative artificial intelligence (genAI) is intense. Tech-savvy consumers are quickly jumping on the genAI bandwagon for activities ranging from search to entertainment to travel and meal planning, however, it’s a different reality for the enterprise.

Despite the hype, the truth is that we’re still a couple of years away from full enterprise implementation of genAI apps like Bard, ChatGPT, DALL-E and others that are gaining traction. Top reasons for the delay include the need for businesses to protect proprietary information, their heightened cybersecurity concerns and a lack of AI skills among the workforce. These concerns have led to some of the world’s largest organizations banning the use of genAI within the work environment.

However, genAI done right will bring enormous benefits to the workplace making it unrealistic to believe that companies will not eventually implement the technology. In fact, many industries have already applied genAI to roles such as customer service and workplace collaboration and are already seeing a positive impact on employee workloads, productivity and efficiency. 

As organizations address data privacy, security issues and the workforce upskills on the expertise needed to use genAI tools and applications, enterprise IT professionals can take several steps now to prepare for future implementations. This article looks at four critical areas businesses need to address when preparing to add genAI to their workflows.

Step 1: Prepare the data 

One of the most critical steps to help organizations prepare is to determine their data liquidity. Data liquidity is the ease with which data flows across different systems, processes and is accessed by individuals within an organization. Achieving data liquidity ensures that the necessary data is accessible to the teams responsible for uploading it into the large language models (LLM) when it’s time to train genAI systems for their respective business functions.

Data liquidity not only enables seamless integration between different data sources and systems within the enterprise IT stack, but it also allows for the consolidation of data from various databases, applications or APIs that might be needed for training. Smooth data integration makes it easier to combine diverse datasets, extract relevant features and compile a comprehensive dataset for the genAI model.

Once the model is up and running, data liquidity enables real-time or nearly real-time updates to the data used by genAI-enabled applications. This ensures the information fed into the model is up to date, which is particularly important in environments in which data evolves rapidly such as financial markets, IoT sensor data or social media feeds. With access to fresh, relevant data, the genAI model can provide more relevant analysis and insight.

Step 2: Adopt a multi-cloud strategy

A multi-cloud approach can both reduce risk and save on costs, enterprises considering genAI should adopt a multi-cloud approach to minimize the impact of potential service outages or other disruptions that could occur with an on-premise implementation or a single cloud provider. This diversification helps to maintain essential business continuity and ensures that data and applications remain accessible. This strategy also allows enterprises to take advantage of competitive pricing models offered by various providers, optimize costs based on specific workload requirements and provide flexibility in choosing the most cost-effective cloud services.

A one-size-fits-all approach doesn’t fly in today’s complex cloud landscape. Every cloud service provider offers its own unique capabilities, features and pricing models. Similarly, each enterprise has its own distinctive workload requirements. Drilling down, different business units and the applications they use might have specific requirements that are better addressed by one vendor over another. In healthcare, for example, scalable data analytics might outweigh specialized genAI services for an organization that needs to move and analyze massive amounts of patient data for research purposes, identifying patterns, predicting disease outcomes and optimizing treatment plans.

A multi-cloud approach to genAI is also beneficial in enterprise environments as they often demand significant processing power for fine-tuned learning as well as operations. Enterprises can scale up or roll back   their infrastructure requirements as needed without the burden of managing large on-premises data centers. A multi-cloud approach also enables seamless integration of genAI models into existing applications, as organizations can choose the most suitable cloud services for each. This can streamline the implementation process and controls costs by using the most cost-effective cloud resources for each specific task.

Step 3: Begin citizen development programs

Unfortunately, today’s demand for skilled data scientists and AI specialists exceeds supply. To address this gap, organizations can prepare for the future by equipping their existing workforce with low-code and no-code tools. So-called “citizen developers” are employees who might not have a formal technical background but possess domain knowledge and a willingness to learn. Low-code and no-code platforms provide visual interfaces and pre-built components to simplify the development process, enabling nearly anyone with the ability to create and deploy applications without needing to write extensive code or have deep technical expertise.

These platforms enhance productivity enabling business units to quickly prototype, test and deploy genAI-enabled applications while reducing the reliance on scarce technical resources and enabling faster innovation and decision-making.

It’s important to note, however, that while low-code and no-code platforms can bridge the talent gap, they of course don’t replace the need for technical experts in the development of genAI-enabled applications. The key is finding the right balanced staffing approach in which citizen developers work in collaboration with AI specialists to ensure the accuracy, reliability and ethical use of genAI applications.

Step 4: improve processes 

Natural language processing (NLP) is one form of genAI that has already started to gain traction within the enterprise. Companies that haven’t yet determined how genAI can aid their business should start by assessing possible opportunities in operational functions. For example, a genAI-powered NLP could triage IT support request emails. IT teams can use NLP to automate the initial sorting and categorizing of incoming support requests, analyzing the content, extracting key information and classifying the emails into relevant categories or assigning them to the appropriate team. This streamlines ticket management and improves response times.

Other popular use cases for NLP include automating customer service in retail, evaluating user-generated content on social media platforms and analyzing financial data to manage risk.

Widespread enterprise implementation of genAI can be accelerated by taking steps to ensure data liquidity, adopt a multi-cloud strategy, enlist citizen developers and leverage NLP.  By addressing these areas now, businesses can position themselves to embrace genAI and unlock its transformative potential.