Beyond the hype: How enterprises can harness the power of AI

by Silvia Gomez Dominguez

From traditional predictive intelligence to today’s chatbots, there’s no doubt that AI solutions are one of the biggest trends in business, with innovative solutions now being applied in day-to-day operations. The challenge for enterprises, though, is in how to look beyond the hype to discover where the real utility is for AI within their organisations. Then, once a problem has been identified that AI could solve for, how can businesses integrate and operationalise AI to realise benefits? We interviewed Blackdot Director Silvia Gomez Dominguez to learn more about how enterprises can get the most out of their AI capabilities.


Silvia, you’ve worked across a number of organisations that have dabbled in AI experiments to varying degrees of success. In your view, what is the case for utilising AI in an enterprise setting?
 
There are a range of benefits AI can offer – most importantly, it serves to improve customer engagement, and also has the potential to identify opportunities to either grow revenue or reduce cost. At end of the day, it’s all about learning from your customers’ past behaviours and using that data to prioritise where to focus your efforts next. We’ve seen businesses leap ahead of their competitors by using AI – when it works well, it boosts performance across the entire business.
 
What is an example of an organisation that has leveraged AI successfully in recent times?
 
When people talk about using AI successfully, what this usually means is using data to reduce customers’ frustrations. Companies as varied as Afterpay, Woolworths and Youi Insurance – to name a few – have harnessed AI to enrich their data, improve their decision-making processes, and also co-share benefits with their loyal customers. If we oversimplify for a second, AI is all about interpreting customers’ past behaviours. Once the model knows how each persona ‘behaves’, it’s then just a matter of applying that knowledge to similar personas. And the more the model is iterated, the more accurate it becomes!
 
How do you figure out which business problems AI can solve for?
 
You can add an AI model onto almost any problem. If it involves data, AI can help. Traditionally, organisations have used AI to reduce their marketing investment – moving away from macro ATL campaigns, to micro and more targeted campaigns which outperformed the former, translating into a significant uplift in profitability and ROI. Then, organisations started to look at AI in a broader way, so that they could provide more value using third party data sources and unstructured sets of information. Nowadays, the ultimate benefit of using AI actually resides in improving the customer experience while optimising cost.
 
How do you set up AI solutions for success?
 
Before considering AI as a solution, it’s vital to ensure you’re able to set up the model to gain feedback and close the loop in between iterations so that the machine learns. Unless you’re able to operationalise and embed the model within your current operations, it’s complicated to close the loop and feed it back into model. I recommend starting with an MVP – create a minimum viable product of your AI solution quickly, then launch it, and iterate. You must train AI, and this only comes with test and learn.
Also, make sure that you prioritise data cleansing for the purpose of your specific use case – it’s not just cleaning the data for the sake of it, it’s about prioritising what you’re going to use it for, then establishing how you’re going to use the data so you can apply a model on top of that. 
 
What about the change aspect, any tips on how to get buy-in when introducing the new model?
 
Absolutely, there are a few tips worth bearing in mind:

  • Be very clear about the short term benefits and the long term evolution – socialise the short term benefits well, but have a plan in place for the long term benefits.
  • Ensure there’s collaboration across all relevant areas of the business, so that the customer benefits are considered more broadly.
  • Partner specifically with the technology team, as they need to be across and onboard with what you’re doing if you want to embed it long term.
 
So, if there’s one thing to remember about AI, what would it be?
 
Whatever the problem is – there’s intelligence we can put on top of it, using the data and AI to approach it, subvert it and go around it. By seeing AI as a problem-solving tool rather than an end goal in itself, organisations will get maximum benefit out of their investment and their customers will have a tangibly improved experience.