GenAI Meets the Chief Transformation Officer GenAI Meets the Chief Transformation Officer

GenAI Meets the Chief Transformation Officer

This is the fourth story in the “Our GenAI Journey” series.

There’s nothing speedy about transformation.

To manage it effectively requires strategy, diligence, and patience. As part of this work, Chief Transformation Officers (CTOs), of which I’m one, identify short-term and long-term goals for improvement and growth and then manage the execution to measure each. It is deliberate work.

But sometimes external forces, especially those in the digital realm, can disrupt the rhythm of these thoughtful strategies. Think of the rapid rise and impact of mobile computing or cloud. In both cases, monumental decisions around existing infrastructure and software investment, management, and security, to name a few, had to be reconsidered quickly.

To succeed in times of disruption, the transformation officer must keep a firm hand on the tiller and remain committed to the guiding principles and best practices of the role: they must adapt to the innovation, identify new objectives, and then balance investment against those objectives.

It’s not for the faint of heart. On the role and challenges facing the CTO, a recent story from McKinsey & Company, noted: “In such a dynamic business environment, focusing on new ways of working, new capabilities, and new technologies is the way forward. Yet transformations are not easy to get right. Research by McKinsey has long documented that enterprise-wide transformation is difficult, with less than a third of transformations reaching their goals to improve organizational performance and sustain these improvements over time.”

The GenAI Transformation

Generative AI (GenAI) is definitely one of the latest disruptions driving transformations that people need to ‘get right.’ Whether arriving at the proper approach to infrastructure (cloud vs. on-premises), selecting the most effective tools, settling on public vs. homegrown large language models, or a host of other concerns – ‘regulating’ the flow of GenAI, much like the step-down electrical transformer, has become paramount for the transformation officer.

That’s why one of my primary focuses has been on ensuring we have our GenAI priorities in place across the business. That is to say, we must recognize that the individual organizations within our company have their own distinct set of priorities that are based on their respective goals and objectives.

On the engineering side, people like Bharti Patel and her team at Hitachi Vantara are focusing on how to accelerate the innovation within the product set and bring new products to market that incorporate GenAI. Whereas, Prem Balasubramanian, with Hitachi Digital Services, is focused on leveraging the company’s professional services and consulting to take his customers through their own GenAI journeys.

Because each group has distinct priorities and operates at different rhythms, we know the innovative process is also calibrated differently and against the respective business goals of each.

GenAI Investment: from Production to Privacy

Once priorities are set and captured, the real work begins. For us at Hitachi, we determined that groups looking for investment in GenAI be required to measure results in terms of revenue growth or cost optimization. For example, in the case of product development, if we determined that the integration of GenAI would enable the product to be easier to use or faster, we should be able to charge a premium. That being said, it should be noted that GenAI investment comes with a big caveat, one like I’ve not seen before. That is, the technology ramped up so dramatically over the last year, from new innovations, updates and tools to new regulations and guidance, that much of the strategizing and investment has had to be achieved with last year’s budgets.

To paraphrase Bharti from the second story in this series, most organizations have had to, build the plane as they flew it. For many, that has meant there has been little to no incremental budget with which to invest in GenAI programs. This has pressed companies to be as creative and deliberate as possible.

In addition, we have found that the short-term return for GenAI, from a revenue growth standpoint, is just not there. That means, I should not spend $1,000,000 today and expect a $2,000,000 return in the next six months. The actual return on investment period is much longer with this technology and it may take organizations some time before getting that equation close to a predictable state.

That’s not to say there may not be some low-hanging opportunities to get some return from a “cost” standpoint. For example, we rolled out Microsoft Copilot (previously Bing Chat Enterprise) to all employees at the end of November last year. The LLM chatbot offers GenAI capabilities like text & code generation, efficient email processing, analysis, and enhanced research capabilities. So far, an estimated 1,000 employees are using Copilot at least three days per week, resulting in a savings of about two hours a day per employee. That adds up to an estimated time savings of about 6,000 hours per week – 1,000 x 3 x 2/week. From a cost standpoint, that’s a big win.

From a privacy and security standpoint, few innovations over the years have required as much attention in this area as GenAI. For us, among other things, it has meant advancing our cyber security & privacy team with new responsibilities to evaluate, approve or deny the development or use of GenAI tools and services. It has meant ensuring our work and developers are cognizant of the risks of potential bias in models. We must insist on maintaining the highest level of integrity in the data, the lifeblood of AI, across investment, development, procurement, and use.

To that end, we have instilled a privacy-centric approach to investment. This involves screening tools and development for everything from internal data privacy guidelines to external legal requirements, which are equally as important. For example, the inadvertent use of copyrighted content is also a real risk. Just ask The New York Times, which used OpenAI and Microsoft late last year for copyright infringement. The Times also followed other media companies, including the BBC, CNN, and Reuters, in making it impossible for OpenAI to scrape its web pages for use in AI training processes.

Whether you intend to use an online opensource GenAI tool or service in which you upload data into the cloud, or you’re developing products using opensource LLMs by pulling in trained models, it is imperative to be critical of the voracity of the security options, as well as the integrity of the incoming data. It comes down to protecting your intellectual property (IP), while simultaneously protecting yourself from inadvertently using someone else's.

This is true for volumes of data down to code. Imagine a developer is writing a productivity improvement for an application and pulls in a code snippet automatically. It sounds nice to everyone because it’s a real time-saver. But there may be a serious risk that the code snippet itself is copyrighted. Someone may have just picked it up and inserted it. That is a problem.

Conducting from the Front Row

From our internal product and policy development at Hitachi, to our customer-facing portfolio and partner solutions, we have been surgical about how and where we invest, as well as how we govern.

As the Chief Transformation Officer, I have a front row seat to these challenges. From this vantage point I can orchestrate the programs and plans to ensure our various and distinct groups progress in concert, with little to no overlap, leveraging and templating what works wherever we can across organizations.

In fact, one of the first things we did to avoid overlap, as well as set a course for the future, was to create a cross-functional team of people with different disciplines across HR, Legal, etc. We’ve been meeting regularly since last summer to discuss successes and failures and to make each other aware of the various projects being considered. In addition to keeping awareness high, we leveraged the brain trust of the cross-functional team to settle on the company’s direction for GenAI and the overriding goals to optimize costs and grow revenue.

The GenAI Horizon

Although this may be a contrarian view, GenAI is not the AI nirvana some would like to think; it is a main component in the broader realm of AI. But it is game-changing. In many ways it has democratized the art of the possible. For our part at Hitachi, we’ve spent decades developing with AI and GenAI. I personally have been working with the technologies since the 1990’s. As the President of Hitachi Digital, Gajen Kandiah, has said many times, GenAI can and should be managed aggressively but thoughtfully, with clear guardrails. And as Prem said, “everyone is learning as they go.” As CTO, I couldn’t agree more with these sentiments. I am optimistic about the future of this innovation and all its potential. And I look forward to continuing orchestrating this transformation for our long-term success.

Santhosh Sreemushta is Chief Transformation Officer, Hitachi Digital.

Our GenAI Journey Series

Part 1. Introducing Our GenAI Journey
Part 2. Tracing Our First GenAI Steps
Part 3. Taking the GenAI Learning Journey with Customers