top of page
Search

Boost Business Value: Leveraging Strategic Insight to Chart a Route to Generative AI

Few tech advances have captured the public imagination the way that ChatGPT has: Just two months after its launch, the LLM-based chatbot had 100 million users, making it the fastest-growing consumer application ever.


Recently, I showcased the LLM (large language model) foundation models that underlie generative AI solutions like ChatGPT. This technology is rapidly gaining significance as ever more business leaders recognize its potential to radically increase performance – transforming business, science, and society as a whole.


The current drive to adopt generative AI is understandable. But implementing it in a corporate setting is complex and requires painstaking planning and careful consideration. This month, I’d like to share some insights into how to put generative AI to work for your business.


Generative AI and Business Value


If you’re a CXO thinking about introducing generative AI, the first question you must ask yourself is this: How will generative AI contribute to business value? To find the answer, you need to understand not only the tech, but also the business opportunities it can offer.


Gaining this understanding initially involves identifying use cases. Additionally, you should consult closely with experts from the various areas of your organization’s business and listen to their ideas about how best to integrate the tech into existing business processes. And, of course, you need to make sure that implementing the tech is in line with your overarching business strategy.


Is Your Current Technology Stack Ready for Generative AI?


Having addressed the business aspects, it’s time to zoom in on the technology. Here, you first have to assess the costs associated with the tech and ensure sustainable energy consumption. Then, you should compare the costs and benefits of generative AI with those of other AI and analytical approaches, which may be a better fit for some use cases – and also significantly less cost-intensive.


Because LLMs and generative AI call for hefty computing power, it’s vital to assess your organization’s readiness in four key areas: technical infrastructure, architecture, operating model, and governance structure. Let’s look at two of these areas in more detail.


Quality and Accuracy: The Pivotal Role of Your Data Infrastructure


To enable your organization to leverage generative AI effectively and efficiently, you must assess your data infrastructure – and update it, if necessary. Generative AI models not only consume large amounts of energy; to deliver accurate outputs, they also require vast volumes of high-quality training data. But many businesses simply can’t provide this. In fact, a recent survey discovered that 77% of companies have issues with data quality.


Given the importance of these factors, you need to examine whether your current data pipeline can deliver the high levels of data availability and quality needed to make generative AI a success. If that’s not the case, you should introduce standards to ensure high-quality data via data cleansing and data governance.


What’s more, you have to determine whether your current data infrastructure can be integrated seamlessly with generative AI platforms and tools. Here, you may find that a modern, cloud-based enterprise data platform is necessary to help break down existing silos within your organization.


Gaining and Maintaining Trust: Governance


Wherever AI solutions are deployed, customer trust is paramount. That’s why you should establish an effective governance structure for managing generative AI. Governance of this kind should enable data security while regulating outputs and their correctness.


Your governance executive team should implement clearly defined data-security and compliance guidelines covering ethical, legal, and technical aspects. Because generative AI systems often produce plausible-sounding but factually inaccurate responses, the team should make sure that all outputs are assessed for accuracy, appropriateness, and usefulness.


And, finally, your governance team should establish place policies and controls to detect biased outputs and deal with these in line with your organization’s policies and applicable legal requirements.


But Where Do I Begin?


So, how should you go about tackling a generative-AI initiative? Drawing on my experience from real-world projects of this kind, I generally recommend the following approach. Whatever you do, don’t dive straight in; test before you invest. In generative AI initiatives, start small, with manageable pilot projects geared to testing the feasibility of generative AI in your business context.


Three Proven Approaches to Implementing Generative AI


With the findings from these pilot projects under your belt, you can then move on to actual implementation. There are three distinct approaches to using generative AI:

  • Off-the-shelf

  • Prompt engineering

  • Custom


The off-the-shelf approach involves using existing foundational AI models directly and entering appropriate input prompts. For example, you could use a pretrained model to create job descriptions for your organization or to suggest subject lines for marketing emails.


Prompt engineering, by contrast, entails programming software and connecting it to your LLM in order to leverage foundational models, making for better responses. And, finally, the custom approach involves tuning existing models by adding proprietary data or modifying their architecture, thereby significantly changing their behavior. While costly, this approach offers the highest degree of flexibility.


Redefining Traditional Tasks


Harnessing the power of generative AI offers companies a host of opportunities for generating business value. But if CXOs are to integrate the tech into their organization, they must take a number of critical steps.


Charting a route to generative IT value entails reinventing traditional work. Business leaders must spearhead this change by redesigning jobs and tasks, and reskilling people. And they must start now. To integrate generative AI successfully and drive business value effectively, you need a holistic approach that encompasses business understanding, technology readiness, data infrastructure, governance, and strategic pilot projects.


Questions? Opinions? Insights?


Want to find out how exactly generative AI could benefit your organization? Then, feel free to reach out to me to arrange to a time to talk. And, as ever, if you have ideas and experiences of your own around this month’s featured tech, you’re more than welcome to share them in the comments section.

 

0 views0 comments

Comments


bottom of page