The Futurist: AI's evolution in business, from understanding to experimentation to implementation

As a futurist for Tata Consultancy Services, I spend my time looking across domains including science, technology, economics, geopolitics, the environment, and others to help our large enterprise clients better understand our rapidly evolving future.

I also spend a lot of time studying history, another important factor that often helps inform our future. In that context, this month is an interesting time to reflect upon AI. We sit between the second anniversary of OpenAI’s release of ChatGPT and the end of the year, so it’s a good opportunity to look back on what’s happened over the past two years and look ahead to what we might see in 2025.

As I think about these past, present, and future macro trends, I generally break it down this way:

·      2023: The Year of Understanding

·      2024: The Year of Experimentation

·      2025: The Year of Implementation

Each of these phases has presented unique business opportunities and obstacles, driving transformative changes across industries. Let’s look at each of these stages and the key trends and insights that mark the journey of AI from a novel concept to a mainstream business solution.

2023: The Year of Understanding

As I reflect back on the early days of 2023, I remember how businesses had a collective “freak out” moment when they saw generative AI tools like ChatGPT, DALL-E, and Bard capture the public imagination, introducing AI’s ability to generate text, images, and audio with unprecedented creativity and efficiency.

Businesses were unprepared for the implications of these powerful tools and they generally responded by blocking employees from using them on company networks. This is why I encourage businesses to stop trying to predict the future, and that they should rehearse it instead, preparing for a range of possibilities. (That’s a topic for another article.)

However, the head-in-the-sand response by businesses didn’t last long, as they quickly realized they needed to embrace this groundbreaking new technology or risk being left behind. That said, this phase was largely about exploration rather than execution.

Opportunities:

1. Awareness Building: Companies invested in understanding AI’s potential and identifying areas where it could deliver value. This foundational knowledge set the stage for subsequent experimentation.
2. Early Use Cases: Generative AI found initial applications in marketing, customer service, and content creation, offering glimpses of its transformative potential.
3. Ecosystem Development: Partnerships between AI providers and businesses began shaping ecosystems for customized AI solutions in sectors like healthcare and finance.

Obstacles:

1. Accuracy and Reliability: Generative AI models, while impressive, were prone to inaccuracies and "hallucinations," where they generated incorrect or misleading outputs. There were some high-profile and embarrassing examples of these shortcomings, which undermined trust in their capabilities and highlighted the need for human oversight.
2. Data Privacy and IP Concerns: Nothing keeps an enterprise CIO or Chief Security Officer up at night like concerns over data privacy and intellectual property breaches. They were right to quickly identify the potential for these issues with all the major LLMs, and they worked hard to ensure their businesses weren’t affected. This emerged as one of the key barriers to early adoption.
3. Workforce Hesitancy: While some employees immediately saw the potential for AI to improve their productivity and job satisfaction, many others were skeptical about AI, fearing job displacement or skill redundancy. Based on my interactions with non-management workers, a lot of this hesitancy resulted from their lack of exposure to generative AI. They simply hadn’t experimented with this new technology and were intimidated by it. We can blame that on the human condition: We’re a stubborn species that’s resistant to change. Nonetheless, this necessitated focused efforts on change management and education, which will be ongoing in the coming years.

2024: The Year of Experimentation

As I look back over this year, I see that many businesses transitioned from understanding to experimenting with AI’s capabilities. We began to see pilot projects, proofs of concept, and sector-specific trials aimed at testing AI’s applicability and effectiveness. I also saw more widespread adoption from employees leveraging AI to help them be more productive.

Opportunities:

1. Sector-Specific Trials: AI-driven tools were piloted across diverse industries. For instance, generative AI enhanced medical imaging and synthetic data generation in healthcare, while creative tools revolutionized marketing campaigns. As a former marketer myself, I have the pleasure of being connected to a lot of brilliant marketing executives on LinkedIn, and it’s been fascinating to see the creative and innovative things they’re doing with AI.
2. Process Improvement and Analytics: Companies leveraged AI to streamline processes across IT, HR and customer support, improve analytics in manufacturing and the supply chain, and improve financial planning and analytics. These experiments demonstrated AI’s ability to augment decision-making processes.
3. Multimodal Capabilities: Integration of text, images, and audio opened avenues for immersive user experiences and advanced analytics.

Obstacles:

1. Cost of Implementation: High costs associated with AI development and deployment led some companies to scale back or delay their initiatives.
2. Talent Shortage and Skills Gap: This year’s rapid adoption of AI has highlighted the shortage of both skilled professionals who can develop, manage, and integrate AI systems, as well as the difficulty with upskilling existing employees to work alongside AI technologies. This is an important one for us all to remember and embrace. Just because you aren’t an expert in data science or machine learning doesn’t mean you can’t become an AI practitioner. Prompt engineering, critical thinking, domain knowledge and a host of other non-technical skills will be critical to your and your business’ success in AI. Developing these skills will help you grow your career and avoid being automated out of a job. One of my favorite quotes on this subject is, “AI won’t take your job, people who use AI will.”
3. Data Quality: Data was another bottleneck in AI experimentation, as poor-quality or biased datasets undermined model effectiveness, while the sourcing, cleaning, and labeling of training data significantly increased project delays and costs.

Despite these challenges, 2024’s experimentation phase provided invaluable insights, laying the groundwork for broader AI adoption in 2025.

2025: The Year of Implementation

2025 looks to be the year when AI moves from experimentation to large-scale implementation. Businesses across industries are pursuing opportunities to integrate AI into their core operations, driving tangible value and competitive advantage.

Opportunities:

1. Hyper-Personalization: AI is enabling businesses to deliver highly tailored customer experiences. Rapid advancements in AI models and algorithms will allow businesses to create hyper-personalized experiences in industries from healthcare to education to e-commerce.   
2. Intelligent Process Automation: AI-driven automation will transform workflows by handling repetitive tasks such as data entry, invoicing, and documentation. If applied effectively, this will free up employees to work on more strategic initiatives that not only make them more valuable and productive but improve their job satisfaction as well.
3. Scientific Advancements: In an area I’m especially excited about, generative AI is accelerating research and development, particularly in fields like drug discovery, where it identifies patterns in massive datasets to generate new hypotheses. It remains to be seen if 2025 results in truly groundbreaking discoveries that improve societal health and wellness, but it’s on the near-term horizon.

Obstacles:

1. Regulatory Ambiguity: As we usher in a new presidential administration next January, it’s unclear how the political and regulatory climate around AI will change. Early indications point to a more permissive stance in the US. That said, businesses will still need to ensure compliance with other regulations like the EU AI Act, creating complexity as global companies roll out new AI capabilities.
2. Cybersecurity Threats: The dual use of AI for both cybersecurity and cyberattacks will pose a significant challenge for businesses, requiring robust defense mechanisms to safeguard sensitive data.
3. Integration Hurdles: As organizations (particularly large enterprise businesses) work to scale their 2024 pilot projects, many will struggle to integrate AI tools with deeply embedded legacy systems, resulting in stalled projects and budget overruns.  

After two years, one might think I’m beyond having my mind blown by AI. But that hasn’t happened yet. In fact, the blistering pace of AI development and capabilities, and the subsequent velocity of businesses working to implement AI continue to be breathtaking. I don’t see either of these things slowing down anytime soon.