Generative AI Readiness: What Does Good Look Like?

Article by Bethan Williams, Global Portfolio Lead – Applications and Data Consulting at Dell Technologies.

Organisations of all sizes in virtually all industries want to infuse the power of generative AI (GenAI) into their operations. How does an organisation prepare to take full advantage of generative AI across functions, departments and business units? What are the most important capabilities to build up or acquire?

To Achieve High GenAI Readiness, You Need a Framework

To help you be intentional about your generative AI readiness, we’ve defined a framework that covers six dimensions of readiness:

1. Strategy and Governance
2. Data Management
3. AI Models
4. Platform Technology and Operations
5. People, Skills and Organisation
6. Adoption and Adaptation

The following are some highlights of what higher levels of readiness look like for each of these dimensions. Note that these are descriptions of future states for these dimensions, snapshots of your GenAI destination, so to speak.

Most organisations will implement many GenAI projects at the same time they are progressing along these dimensions, and the lessons learned from those early projects will help inform the readiness improvement efforts.

Drive GenAI Strategy with Business Requirements, Use Cases and Clear Governance

In an organisation with a high degree of GenAI readiness, business and IT leaders collaborate to set clear objectives aligned to business priorities and actively manage a GenAI project pipeline.

Given the exceptional opportunities for innovation and optimisation GenAI brings, it is more important than ever for organisations to achieve consensus in their transformation strategy. Starting with a focused set of strategy workshops, including all stakeholders who will be involved in this transformation, ensures all voices are heard, facilitates the path to agreement and gives everyone a solid vision of the future state of the organisation and how to get there.

It’s vital to gain a clear view of the use cases that are most important for the business. Organisations often struggle with prioritisation, as potential GenAI use cases extend into every corner of the enterprise. As part of our Professional Services for Generative AI, Dell Technologies has created a use case prioritisation tool so business, IT and finance professionals can identify, analyse and prioritise use cases according to business value and technical feasibility.

With new use cases comes potential risk, making it especially important for organisations to have effective oversight of all GenAI projects. This ensures compliance with regulations, risk management guidelines and evolving ethical considerations.

Get Your Data House in Order

Many organisations start their generative AI journey using pre-trained models, which require access to an organisation’s data to provide the context needed for successful implementation of GenAI use cases. Whether that data is provided via model tuning or augmentation (e.g., Retrieval Augmented Generation or RAG), delivering good data to the model in a timely manner becomes key to GenAI success.

As such, a high-readiness organisation prioritises scalable data management as a key enabler for GenAI, coordinating discovery, acquisition and curation of data. Business analysts and stakeholders should have access to an easy-to-use catalog of enterprise data resources.

With data management now in focus, organisations can ensure data is clean prior to use, reducing errors and bias and preventing exposure of proprietary information. A good indication of maturity is the use of data models to support both structured and unstructured data, simple integrations, automated transformations and pipelines.

Match the Model to the Use Case and Continuously Monitor Performance

Given the costly, time-consuming and expertise-intensive nature of training a model, many organisations will choose to use techniques such as RAG, prompt-engineering or fine-tuning of a pre-trained model to quickly realise value from GenAI.

The number of choices available to customers when selecting pre-trained models is growing daily, which presents new challenges and new opportunities. Key factors in model selection should include user experience, operations, fairness and privacy, and security.

Selecting the right model is just the start. A high-readiness organisation establishes processes for evaluating the performance of its chosen generative AI models, regularly tuning model parameters to optimise effectiveness. Organisations should frequently assess models for safety, fairness, accuracy and compliance.

Build a Solid Technology and Operational Foundation

Once an organisation selects use cases and models, they need a trusted platform to implement and run them. The mature organisation will utilise a GenAI technology stack appropriate to their use cases, security and data constraints, and ensure these technologies are standardised across the organisation and priority use cases. AI data is seamlessly integrated with multiple data sources.

Scalable data management is key to GenAI success, so highly mature organisations will have a GenAI-ready data management architecture such as Dell’s data lakehouse for analytics, with advanced analytics tools.

Level Up Skills and Organisation

People with AI skills are well positioned to embrace GenAI. However, there are new skills needed beyond those required for traditional AI. A high-readiness GenAI organisation provides training for specialists on platforms and tools, architecture, data engineering and the like. End users learn data analytics principles and how to construct effective prompts. This is supplemented with new support and operations teams dedicated to generative AI.

Manage Adoption and Adaptation

An organisation at a high level of GenAI readiness has a clear understanding of where and how generative AI can add value. The initial strategy sessions help create that early view, but this is not a static space. Business and IT must continue to work together to integrate GenAI into new initiatives.

Continuous improvement within GenAI should be standard practice for organisations and can be achieved in a number of ways. Teams can capture human and automated feedback from model outputs and incorporate lessons learned into model training, guardrails and information retrieval.

These organisations integrate automated compliance with corporate policy, data privacy and government regulations into development and deployment processes.

Embark on Short-term GenAI Opportunities and Advance GenAI Readiness

As an organisation moves to higher readiness levels, the opportunities for leveraging the benefits of generative AI increase in number and business impact.

But don’t think you need to wait until the readiness dimensions reach a certain level to begin applying GenAI to key use cases. You can and should begin with shorter-term, tactical projects that can provide efficiencies and financial benefits today.

If you’re looking to apply GenAI best practices, Dell Consulting Services can help in many ways. A great place to start is a Generative AI Accelerator Workshop, a half-day interactive strategic session with business and IT leaders to assess your organisation’s GenAI readiness.