Pavan Bedadala, Senior Director of Product Management at Commvault, discusses how it’s critical for businesses to harness the right strategy at the beginning of their AI journey.
With the scorching pace of advances in their AI journey, businesses have been thrust into reactive mode.
As AI continues to dominate the business conversation, CEOs and board directors increasingly demand that every department figure out how to adopt the technology. Too often, they seek immediate answers. Why?
- They want to see business units automate more processes to help, among other things, drive better margins
- They want to quickly dazzle customers with new applications that provide an individualised experience for each user to help build more brand loyalty
- And if it’s a public company, it wants to go into its next investor call armed with a multitude of AI-related projects it can tout to Wall Street to increase shareholder confidence
But a quick-win mindset doesn’t work for AI. Without taking the time to analyse, understand, and clearly think through why and how the tech can benefit their specific business, companies risk wasting millions of dollars investing in systems that do nothing to improve efficiency or give employees the power of predictive intelligence.
Worse yet, starting too many half-baked projects all at once could create chaos and actually harm the business.
That’s why the first 60 days are the most important when launching an AI strategy that can support the business for the next decade.
- It’s when company leaders should highlight specific business objectives or problems and consider how or whether AI can help
- It’s when they should implement plans to democratize data access so that, over time, all employees can have access to the best information and automation to make them more efficient and effective at their jobs—all without jeopardising IT security
- And it’s when business leaders should commit to foundational processes that will ensure that the AI systems, once tested and proven in controlled environments, can scale to the whole organisation without diminishing the quality of the results
In other words, businesses mustn’t try to shortcut the AI adoption process. By setting the right goals, creating the proper data strategy and putting the right oversight structure in place, companies can embark on their AI journey with confidence.
Set the right goals
Often, in their haste to start seeing the benefits of AI, companies skip the goal-setting stage.
Business leaders might be reluctant to spend valuable time reflecting on how and why to make an AI investment. Instead, under pressure from above, their heads are clogged with a million to-do items that need immediate attention.
But when done right, most business leaders quickly realise how valuable accurate strategic planning and goal setting is to their future.
- When it comes to your AI journey, it’s not about what applications you are deploying. It’s about the problem you are trying to solve
- Identifying bottlenecks in current operations or pinpointing lower-value tasks that are dominating employee time, for example, provides a starting point for identifying where an AI investment might help
- Most importantly, a systematic approach will help in setting specific goals to track progress against
That means that six months down the line when the CEO demands an update on the millions of dollars spent on AI investments, business leaders can quickly show how the technology is meant to meet the business’s larger goals.
Choose the right data in your AI journey
Many companies have been gathering data for decades. Many top executives not involved with IT day to day might be under the impression that the information sets can immediately be used to derive data-driven insights about the business’s future. Given the data sprawl and siloed systems of many enterprises, that’s too seldom the case—but it’s where AI’s data-sorting and winnowing capabilities might be especially useful.
Yet, to ensure that AI models are running properly and generating helpful insights, companies must take the time to set the right data management strategies for their operations.
- They need to audit all the different data sources they already have and start figuring out how to combine them
- They must ensure that the information flowing to the AI models is accurate, clean, and correctly formatted
- And they have to make sure employees are accessing only the data they are credentialed for
Once that’s done, businesses can then start training the models. Most AI systems won’t work perfectly right away. They need to be conditioned through mechanisms like human feedback. Then, once the models run correctly in the test environment, they can be moved to production.
Monitor for the right outcomes
Taking a ‘set it and forget it’ approach with your AI journey can be easy. But the technology is evolving quickly. Almost every day, there’s a new tool or feature to experiment with. So companies must treat each AI model as an ongoing process requiring continual supervision.
- If running correctly, the systems should get more accurate and powerful as they are used more
- But if there are data quality issues, for example, the AI models could produce inaccurate results that may actually harm operations
That’s why it’s critical to constantly monitor all the AI models in operations to ensure they are running properly. Having a running list of each system in production can help make oversight much easier.
It’s obvious that AI has huge potential, but there are also significant risks. That’s why it’s so important to spend the first 60 days getting the strategy right and ensuring the organisation is prepared.
Taking the time to lay a solid foundation now means the business can spend the next decade focused on harnessing the power of future AI advancements.
