It’s Time to Make AI More Human

The advances in artificial intelligence over the past two years have been nothing short of extraordinary. The excitement generated by these advances has led to a massive AI hype cycle that is becoming increasingly divorced from reality. Here are a few contrarian predictions:

• Most companies will not deploy generative AI in a meaningful way in 2024…or 2025.
• Generative AI will create more jobs than it destroys in the next few years.
• The companies that will be most successful in deploying generative AI in their businesses will have narrow, focused use cases that pair technical automation with human expertise to drive results.

Today, members of the board and the executive suite are handing down AI mandates. But the people on the ground implementing these tools are finding it far more difficult to deliver actual results.

To successfully deploy AI, leaders need to do something counterintuitive: make AI more human.

How Is AI Currently Showing Up?

The most sophisticated technology companies in the world have been at the forefront of leveraging automation or AI for some time now. To date, we have largely seen these technologies used as tools to deflect less complex work and to improve the speed and quality of responses for customer service queries, content moderation and generation, and language translation.

Winners in today’s business landscape are companies that can successfully deliver a combination of vertical-specific technology solutions and humans in the loop adept at completing complex customer interactions in real time. Generative AI adoption is gaining traction in all types of industries, but users are often leveraging the same large language models. This includes OpenAI’s GPT (Generative Pre-trained Transformer) series, Google's BERT (Bidirectional Encoder Representations from Transformers), and Facebook's RoBERTa (Robustly optimized BERT approach).

Even with the existence of these AI tools and use cases, there is still hesitation to widely adopt AI in business, especially in the United States. According to report published in November 2023 by the U.S. Census Bureau, 3.8 percent of American businesses use AI to produce goods and services. This could be because many businesses still need to address the myths preventing internal buy-in to AI technology.

Busting AI Myths

If IT decision makers want support from their C-suite on AI, they’ll have to educate them on the truth about AI. Some truths about AI include these:

AI solutions will augment, not replace, talent. These solutions will remove unnecessary manual tasks, increase workflow speed and accuracy, and free up your talent for more meaningful work.

AI won’t immediately solve your business problems. Deploying AI systems requires relentless iteration and an investment of time before value is realized. The most successful companies are focusing their AI initiatives on very specific aspects of their business process.

AI requires human intervention. For businesses to get the most out of investments in AI, they must pair the technology with talented humans. Human team members train, QA, build, deploy and refine AI systems. Additionally humans in the loop step in to handle edge cases or prevent AI systems from making mistakes driven by the risk of hallucination. The integration of human intelligence and AI technologies will define the enterprise of the future.

Leveraging a Human-Driven AI Strategy

Before you implement any AI solution, you need to have a strong data collection strategy. AI is only as effective as the data upon which it is trained. This will likely mean investing in data collection and annotation. Once you have collected and annotated the data, you must invest in a data warehousing solution to ensure all of your valuable company data is in one place. Then, select an AI solution that is specific to your industry and business model. For example, if you want to streamline customer service operations, a generative AI chatbot may be perfect for your business.

As mentioned earlier, training and education will be extremely important for all employees in your business—but especially those who will interact directly with the AI solution. This could mean regular seminars on generative AI, onboarding webinars on automation, and even on-demand content the staff can return to during their free time.

Lastly, develop internal KPIs that can help both IT staff and other business personnel evaluate the AI strategy. With an AI chatbot, this could include checking the accuracy and quality of responses the AI bot provides. It may also help to know how well the chatbot captures data from a client interaction. This could help interactions with future customers. The key is to leverage the expertise of your personnel to ensure the AI solution is delivering results.

AI tools yield the greatest returns when businesses train them on specific data and humans utilize these solutions to ensure efficient, consistent, and secure results. Ultimately, AIs ability to revolutionize the way we work will be up to the humans in charge of implementing, refining, and working with it.

Bryce Maddock is the CEO and cofounder of TaskUs Inc., a provider of outsourced digital services and next-generation customer experience to innovative companies worldwide.