If you are an IT leader, chances are that you’ve caught the generative AI bug by now, and after a period of skepticism, you have started testing its capabilities. You may have also struggled to fully turn the hype into real tangible results. Of course, it’s an IT leader's job to look beyond the hype and ensure proper governance, oversight, and standards before we adopt new technology. Being the CIO of a technology company, I sometimes get to go through this phase faster than my peers in other verticals. After experimenting with the technology over the last few months, I have come to a few conclusions on how generative AI will force IT leaders to change their strategy over time.
Streamlining application portfolios
In the past few years, most CIOs have been binge-buying SaaS applications in the name of digital transformation. There’s nothing wrong with that, especially considering the high long-term cost of building custom applications. Subscriptions have been the logical choice. However, we often bought point solutions that solved only one specific business problem. Generative AI offers a new opportunity.
Combined with an automation platform, generative AI allows you to quickly build solutions without buying a new tool. Based on the assessment of my team, we are hoping to reduce our app portfolio by 5-10% by next year using generative AI capabilities.
At the same time, it may not make sense to build a custom solution every single time. Every supplier is looking at adding generative AI capabilities to their portfolio, so it's a good idea to keep an eye on their roadmap to ensure you aren’t building something that will be available out of the box very soon.
Reimagining user experience
In the last 20-odd years, applications have evolved to become more accessible, running on multi-tenant cloud platforms and becoming mobile-friendly. But their core architecture hasn’t changed. They still leverage a form-based user interface that runs on top of a database that expects structured data. Generative AI challenges the core architecture and experience by thriving on unstructured data and responding with human-like answers.
For example, on Salesforce CRM, it easily takes a sales rep around 15-20 minutes to navigate four or five screens and 15 clicks before being presented with a quote. Why can’t we provide the same functionality through a chat experience that's more intelligent and faster to respond? The same goes for pulling reports and dashboards.
Recently, my team rolled out a conversational way for internal employees to perform basic tasks like booking a meeting, ordering equipment, drafting a sales email by looking at the prospect’s LinkedIn profile, and many more. By combining this with our own automation platform, I expect to expand this capability and turn this into a system of engagement for hundreds of daily tasks.
Simplifying data processing
If you ask an IT leader about one task they hate in every project, I can guarantee that the majority will call out data conversion and cleansing. This is because it involves writing hundreds of lines of code to convert data from various forms to a structure your system can understand. Unfortunately, most powerful analytics tools are only successful if you have figured out data conversion and consolidation.
One of the coolest features of LLM models is that they can make sense of data sitting in different forms. I expect generative AI to change our data analytics strategy by reducing the steps required for data conversion and cleansing.
Revamping team structure
I am part of an automation company, and AI has always been a part of our intelligent automation journey. But it's becoming clear that generative AI will force CIOs to take a second look at their team structure. Similar to building an automation program, you may want to consider creating a horizontal group for AI that supports all aspects of your organization. To that end, I am asking every part of my organization to share and publish their generative AI strategy.
On the skill set side, you might need to prioritize hiring conversational designers over UI developers over time. Similarly, training AI models with prompt engineering is another skill set you may want to invest in. I am also revisiting my partner strategy to fill the talent gaps in the short term. But the biggest challenge will be organizational inertia, and the only way to beat it is to reflect generative AI in individual goals for every part of my team.
Addressing budget implications
No matter how excited we are about the new technology, we all know these are not the best times to ask for an incremental budget. So how do you fund an AI program?
I’d start by looking at my biggest spenders, which are generally CRM and ERP teams. Although you absolutely need the basic capabilities of these tools, there is a real opportunity to leverage automation and generative AI to keep expensive customizations out while keeping the ERP core clean and simple. All of this is possible without impacting the native experience of these tools. At Automation Anywhere, we are using our own Automation Co-Pilot to combine the power of generative AI and automation natively in the ERP experience.
Navigating new challenges
As excited as I am about the potential of AI, there are new challenges to the accountability of decisions made by generative AI. We IT leaders are used to defining rulesets for processes we manage. However, the challenge of controlling application behavior without rules is new to all of us. Keeping that in mind, I highly recommend having a human in the loop for critical actions and constantly reviewing the results.
There is also a much broader need for IT organizations to work closely with your legal team. Generative AI will push the boundaries of our existing corporate policies, and it's imperative to update these policies as we learn from our experience. Consider setting up an AI governance council within your organization to discuss and align on all things AI.
Leading with confidence in a new era
Generative AI is new for everyone, including IT leaders. Still, being the technology leaders for our organization, we’ll be expected to provide a direction, setting the right expectations without compromising on the security and privacy of our customer or employee data. We’ll all make a few mistakes while trying new ideas, and together we’ll learn from the journey and set best practices for the industry.