Balancing AI automation and human insight
From building one of the top ten fastest supercomputers in the world to having robots pick and pack your weekly food shop, the use cases on display at last month’s AI Summit London encompassed everything from functional to futuristic. Something they all had in common, however, was their ability to revolutionize the world as we know it – our daily routines, the way we solve complex challenges, and the way we work.
In this content series, we’ll be tapping into insights from The AI Summit London and exploring the key themes and developments that are influencing and advancing the evolution of Gen AI across the enterprise space. We’ll cover everything from ethics and governance to education and skills – not forgetting AI’s impact on the future of communications, marketing, and journalism.
Here are our four key takeaways from the event to kick things off:
Find the right balance of control and automation
Since the GenAI boom, many businesses have rushed to quickly implement Large Language Models (LLMs), however, there were multiple reminders from speakers that AI does not have to mean full automation and that often a balanced approach is best.
Colin Jarvis, a distinguished architect from OpenAI encouraged organizations to ‘keep humans in the loop,’ saying that it often results in better experiences and outcomes.
Most emerging patterns for success with AI have been across content generation and autonomous assistants, with many well-known brands also exploring options for automating their customer service offerings. Following its acquisition of GamePlanner.AI, Airbnb announced its plans to build the ‘ultimate concierge’, having already tested other use cases including using AI to write review summaries.
John Abel, a technical director from Google Cloud, said the most successful businesses in the future will be those that harness and empower people to do their jobs. The more complex the process is, the more you still want the human involved. The less complex, the more you should automate.
The same rule should apply for marketing and communications strategies. GenAI can assist with drafting press releases and even help respond to potential crises, however, it lacks emotion and brand authenticity meaning human oversight will remain crucial.
Build for the capabilities that are coming
Typically, business are working with out-of-the-box models and applying a small amount of fine-tuning. While this somewhat steers the behavior of the model, you are still stuck with foundational use, so the amount of flexibility is limited.
In the coming months, we can expect to see more organizations taking base models but applying more sophisticated training techniques, meaning they function very differently. This includes the creation of AI models for specific industries, such as banking and healthcare.
Our client Luminance is a world-leading AI platform for lawyers. It’s AI reads and forms a conceptual understanding of legal documents in any language, enhancing and expediting a range of tasks. From taking a first pass review of contracts to highlighting areas of non-compliance, it is a trusted co-pilot for any legal team.
The advice to businesses was don’t build for what is available today, build for the capabilities that are coming. Things are changing so fast that this will be the only way for enterprises to differentiate across industries
Get your data in order
As the ‘new kid on the block’, GenAI is getting a lot of attention. However, AI models can only be as good as the information they are being fed in the form of enterprise data. As every business will know, this data often comes from hundreds of different applications across the organization – the challenge lies in joining up the various components.
This requires a three-step approach to data:
- Relevance – ground AI with your documents, your communications, and your processes.
- Verification – verify AI answers via citations against your data.
- Accuracy – ask better questions to get better, more accurate answers.
Sholthana Begum, Chief Data Officer at the Financial Conduct Authority, said ‘the data and tech are there’ for businesses, their success with Gen AI will be down to how quickly they can sort their data and therefore dictate the pace at which they innovate.
Another of our clients, search AI company Elastic, is helping organizations proactively resolve data observability issues by combining the precision of enterprise search and the intelligence of AI. This allows businesses to get the answers they need from huge amounts of data in milliseconds.
Encourage internal change and create AI advocates
To see success with AI, organizations must get buy in from users – aka their employees. This can be scary for some and exciting for others. Understand that and don’t rush things.
Even if you can automate entire processes and get AI to take over, don’t. If you want to make people feel comfortable and trust the tech they are working with, keep humans in the loop and then gradually let AI step in.
David Lee, Director of Innovation and Growth, and David Norris, Chief Growth Officer, at Holiday Extras showcased examples of how they had unleashed employee AI potential and overcome team challenges. This included giving people permission, making AI fun, creating a safe playground, and giving people the time to innovate alongside their day jobs.
Internal communications strategies will play a fundamental role here. Prioritizing clear and consistent communication not only keeps everyone informed and working toward shared goals, but also results in greater collaboration, adoption, and engagement.
Organizations must ensure they are facilitating any changes around the use and impact of AI tools to employees in a timely and accurate way in order to empower them to make informed decisions and do their jobs effectively.
Our next post will focus on how to develop AI skills within teams across your business, as well as what it means for the future of work. We’ll also be exploring some of the ways that we have been encouraging our team to get to grips with our very own proprietary AI tool, GAIO.tech, which allows our clients to understand what AI is saying about their company and from where it is sourcing that information.
If you’re interested in hearing more about GAIO.tech or would like to speak to our team about the work we are doing with clients in the AI space, feel free to get in touch – hannah.lock@hotwireglobal.com.