workers reviewing tracking system at warehouse factory

How strategic use of an AI maturity model can maximize GenAI ROI

Amid GenAI frenzy, many businesses still face challenges in recognizing its promising implications. Explore the realm of GenAI use cases.

Generative AI (GenAI) is the talk of the town, but while many businesses have been putting new and more complex GenAI solutions on their agenda, many still struggle when it comes to understanding truly promising use cases and the implications for their operations.


Many companies are still wondering: Where and how do we get started with Gen AI.

Nevertheless, experts expect further rapid adoption of GenAI in the next 6 to 24 months. Following the automation of repetitive and manual work already seen today, we expect much more personalization of content at scale in 2024. And within two years, the disruption of existing business models should become more and more common. Accessibility will increase at even lower prices.

GenAI offers many use cases for nearly all business functions.

GenAI can be relevant to nearly all business functions, but use cases need to be specific and tailored to an individual company’s unique setup to allow proper assessment, prioritization and implementation.

In practice, many opportunities pop up fast:

  • GenAI uses in sales and marketing are particularly visible when it comes to automated product description generation. GenAI tools offer opportunities to improve available information for purchasing decisions and to lower the amount of support needed to fill in existing gaps.
  • In human resources and talent acquisition, AI-powered candidate sourcing allows for easier screening of resumes to improve offer rates.
  • For IT and Tech departments, GenAI can help to vastly improve data quality by identifying incorrect, missing or incomplete data in databases.

A simple matrix to classify and prioritize AI projects

To get started, companies can collect relevant use cases at an organizational level, for each member of the C-suite or for each division of its business.

The sheer volume of opportunities makes it necessary to prioritize and identify pilot projects. Here, a simple matrix can help: it allows companies to reflect which GenAI activities offer the best fit with current strategy, available resources in Finance, HR and infrastructure – and the organization’s level of ambition.

Two questions become particularly important:

  1. Does the business or the specific business unit first and foremost require a reduction in cost or is it prioritizing an expansion of its business scope?
  2. To what extent are the specific business processes already supported by standard software or platforms?

The X-axis of this graphic differentiates between projects mostly aiming at reducing cost and those looking for revenue growth. The Y-axis ranks the projects with regards to the degree of standardization of software already available. The higher up a case is, the more it needs individually programmed software; the lower it is, the more it can rely on standard solutions.

Mapping the ideas of a GenAI brainstorming session onto this matrix provides the basis for a structured discussion of the most promising leads and helps with scheduling and determining the next steps.

Hugely innovative and unique projects could secure business growth and justify – but also require – more proprietary software solutions. These projects, often located in the upper right quadrant, might also need a specific type of organizational structure with specific teams, bringing together business, data, analytics and IT functions.

Projects aiming at increasing efficiency make up a large share of use cases at this stage. Some can be realized by employing standard software; however, they may not lead to true differentiation in the longer term. Proprietary solutions may be advantageous in securing a competitive edge over time, but they require own resources, investments and effort and thus can be more difficult to realize. Therefore, some projects in the lower left quadrant may help companies remain competitive in the short- and mid-term and, with standard software improving, more use cases may become lower-hanging fruit.

Not to be fooled, though: Even with standard and emerging software providers integrating more and more AI functionality, users and their data have to fulfill certain prerequisites with regards to data quality, security structure, governance and the like to benefit from this. Whether investing in proprietary solutions or relying on standard software, there is no time to waste.

What needs to happen after classifying GenAI projects

Once companies have identified use cases with the biggest potential, they can start to determine the next steps in their realization of AI projects. These can include more specific discussions about benefits, costs and risks, strategies for how to deal with technical prerequisites for data and infrastructure or needed roles and sourcing strategies for such experts.

Such debates are necessary: Without a doubt, GenAI will continue to improve and to offer immense additional opportunities in many industries and parts of a business. The matrix introduced can continue to facilitate discussions, since the general way of assessing these new projects will stay the same: Future applications can also be located along the two axes. They will also focus either on new revenue opportunities (like new and more personalized after-sales services) or on lower cost (like reducing warranty expenses due to better pattern recognition in damages); and they will come with the need to build unique proprietary solutions (done by massively sought-after experts) or a certain degree of integration into standard software (making preparation today necessary).

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.

Summary

Generative AI can be relevant to nearly all business functions, but GenAI use cases need to be specific and tailored to business strategy.

Related articles

    About this article