Shakespearean man on laptop
Image generated with assistance from DALL-E.

To AI or Not to AI?: Important questions to ask before deployment

McKinsey predicts that generative AI will add “trillions of dollars in value to the global economy” thanks to productivity gains. Given the pace of adoption and projected value, it’s very likely most organizations will start integrating generative AI solutions into their workflows and products in some capacity in the near future. And while the recent global pandemic tested many organization’s ability to be agile and innovate on the fly, this year’s economic headwinds have propelled efficiency to the top of the growth driver pyramid. And with efficiency gains on the menu, it’s no wonder 39% of business leaders will use generative AI every day.

Having said that, it’s important for companies to start by investigating whether adding generative AI to their services adds value to their customers or not. A lot of organizations are too quick to jump on the hype and execute without thinking about the real value to the user. In our engineering and product teams, we are encouraged to ask five times why. The funny thing with generative AI is it's a really different conundrum, and now we are asking 50 times why. Common questions orbit: Why will this add value, how will they use it, what will they ask of it, what outcomes will they get, will it make a difference to their daily work, is it trusted, should we charge for it, will it cost us a lot of money, is it hard to implement, will it be difficult to support? With any new technology, it takes experimentation to fully understand it. And where generative AI is concerned, we’re very much still in the phase of discovering new use cases. So whilst I'd say 'nobody is required to implement it' I would say 'everyone should be experimenting with it'.

Generative AI is simply scaling data and using natural language to extract it. So it's simpler for companies to implement as a plugin. But it can be costly. And tech leaders need to understand these costs before really deeming it a solution to any one problem. 

We’ve heard of use cases where AI chatbots are replacing human customer service departments — a bold move to say the least. Time will tell if this is the right move, but an important question to ask is if the cost of AI to replace a human worker will yield better results, and then on top of that, how can you then utilize those human workers in a way that still builds your business while creating meaningful and valuable work for them? AI should be viewed as a co-pilot, with the expressed role of improving productivity to allow humans to focus on higher-value tasks. What we know for certain is that people using AI the right way will be much more efficient than those who do not. If organizations do not embrace this fact, their reluctance will eventually become a disadvantage to the business in the long run.

What we are seeing in the market is only a beta of a much more sophisticated intelligence engine. These are the early days and we have to continue to learn how to adopt it and put guardrails around it. 

For instance, we don't yet know the privacy and security concerns around using AI and many organizations are scrambling to work out what it means when sensitive data is being put into an external AI engine. When Google Translate first came out, everyone loved it. We all used it to translate simple text, but some companies were using it to translate sensitive documents. Legal teams would eventually advise that this is not a secure application and shouldn’t be used. The same could be said for generative AI!

Critical early steps to org-wide deployment is to establish a working group of early adopters to learn and test new technology. Generative AI is a perfect use case for this model and makes the most sense when this test group — with transparency and consistent communication — can help identify where the technology creates the most value for the business. Only then, when there’s a clear purpose for new technology, should it be deployed at scale.  

At the end of the day, companies and teams need to be flexible, they need to experiment with consideration, and they still need to distill the value of the work they are doing in order to separate hype from using generative AI to develop real solutions for real challenges. 


Lucie Buisson is the chief product officer at Contentsquare, where she leads the product vision and strategy, and co-leads go-to-market. Her team’s mission is to develop innovative products that empower businesses to make the digital world more human. Product addiction, autonomy, and uniqueness are key areas of focus for the product team, as they work on making Contentsquare the world’s leading customer experience optimization platform.