Stop investing so much in A.I. technology

We are concerned that some business leaders may be making existential mistakes.

The proportion of companies adopting AI has plateaued between 50% and 60% over the past few years. Meanwhile, some leaders share their frustration that their AI investments haven’t yet yielded the levels of return they expected. They simply couldn’t scale it.

Are some companies getting ahead of their AI pursuits?

I hope not. Big returns from big companies. Explore innovations that can leverage AI to improve decision-making, speed up processes, automate mundane and repetitive tasks, and push workers into entirely new areas of business.

These are not just tech companies, they are companies in sectors ranging from healthcare to mobility to financial services. And they plan to continue to incorporate AI into everything they do, expanding their advantage over competitors who have not tapped into the potential of AI.

What makes these companies different? We invest a significant portion of our AI investments in capabilities that enable us to do more AI faster with less effort. They are investing in building a true AI factory.

From one-off production to repeatable processes

At the beginning of their AI journey, companies typically run a small number of AI pilots and one-off use cases to tackle business problems. Some companies, with each success, do more of what made those wins happen. For example, we build fully bespoke models and encourage more experimentation in organizational silos.

Continuing to invest in these areas is wise, but doing more of the initial results is not enough to make AI more widely available within your organization. This is the key to realizing maximum her ROI.

As one-off production increased, teams decided to start all development work from scratch, sift through hundreds of raw data sets, write bespoke code, and deploy proprietary development tools and technologies. It also increases the associated complexity and cost.

Scaling AI requires investing in creating reusable assets, platforms, and repeatable processes (a sort of factory). This enables AI teams to build, deploy, and maintain models in less time and with less manual effort.

In our survey, respondents whose organizations benefit most from AI are much more likely to adopt this factory approach. In our work, we have seen many organizations beginning to get this right. Vistra Corp., the largest competitive electric power company in the United States, created an AI factory to standardize the deployment and maintenance of over 400 AI models. With $60 million in savings seen in about a year and $250 million in his EBITDA identified, he is on track to achieve a $300 million roadmap and reduction in greenhouse gas emissions.

Similarly, an Asian financial services company was able to reduce the time to develop new AI applications by more than 50%. Part of that was creating high-quality, ready-to-use data products that could be used on data source systems. A large number of AI applications. The company also standardized supporting data management tools and processes to create sustainable data pipelines, and created assets to standardize and automate time-consuming procedures such as data labeling.

Inside the AI ​​factory

An AI factory is not a physical location, but a framework that enables teams to deliver more AI applications with less time and effort. This includes, for example, giving development teams easy access to the tools and technologies they need through one end-to-end platform that reduces time-consuming and costly integration efforts. Establish a single AI development and deployment playbook and standard set of protocols to bring together company-wide best practices and make them reusable to development teams. We also focus on creating some ready-to-use assets, such as data products and snippets of code, to give your team a head start.

Consider how simply reusing existing code can facilitate some aspects of scale. Until just a few years ago, most AI practitioners had to code every AI solution from scratch. Today, organizations can download state-of-the-art, open-source, pre-trained AI models to use in their application development. For example, a global energy company built a model that accurately predicted customer cancellations by applying just a few lines of code.

At the same time, the new low-code/no-code platform enables employees without extensive AI expertise to develop AI applications. Organizations that benefit most from AI are 1.6 times more likely to use these time-saving tools.

We actively encourage leading organizations to focus on reusability for code that must be written in-house. They encourage and applaud the efforts of data scientists and engineers to contribute their code to a central codebase and make it easily consumable for other applications by other practitioners.

With reusable code available to help non-technical people develop basic AI applications, the most seasoned AI practitioners can focus their expertise on engineering robust, high-value applications for business. can. This increases productivity and enables organizations to do more with the fewest elite AI talent. It also keeps AI practitioners happier because they can focus on the tasks that appeal most to them in their field of expertise. This is key to sticking with the organization.

And, importantly, organizations can redirect talent funding and effort to other new roles needed to build an AI factory. One such role is that of a machine learning engineer. Machine learning engineers are skilled at transforming AI models into enterprise-level operational systems that reliably execute and automate machine learning pipelines, from data ingestion to business insight generation.

Even in this age of chaos and resilience, now is the time to rely on AI. The winning companies double down and prove they can do more AI with less time and less effort to deliver the high returns that technology promises.

Alex Singla is Senior Partner and Global Co-Leader of QuantumBlack, AI by McKinsey, Chicago. Alex Sukharevsky is Senior Partner and Global Co-Leader of his QuantumBlack, AI by McKinsey in London.McKinsey is a partner luckglobal forum.

Our new Weekly Impact Report newsletter examines how ESG news and trends are shaping the roles and responsibilities of today’s executives, and how they can navigate these challenges. Subscribe here.

Source link

Leave a Reply

%d bloggers like this: