Understanding AI and its benefits for business

Companies believe in the promise of AI but struggle with the 'how', says David Wyatt of Databricks

AI - is it really anything new?

Because AI is all the buzz these days but at the same time AI has been around for some time there is confusion about what's really new. AI algorithms have existed for 40-50 years. What's changed is the availability of large data sets and the computing power of the cloud.

However, even then, success in business organisations is still constrained by the complexity of data infrastructure and the collaboration gap between data scientists, data engineers and the business.

Apache Spark is a key recent innovation that made it possible to process big data at much faster speeds and scale machine learning (ML) models that are the backbone of enterprise AI. AI has enormous potential, but until now most businesses have struggled to achieve that potential.

Automating analytics

Two things are happening. In the short term, AI can help predictive analytics projects get results faster through automating some of the steps that would otherwise require human intervention.

However, creating the training data sets and setting up the systems in the right way is still time- and labour-intensive. Speeding up this process - or letting those without specialist skills set those projects up through more support - should help those implementations deliver results faster.

Looking further into the future, I think AI will provide fantastic results in areas like healthcare and life sciences. If you can train AI to spot the difference between cats and dogs in images, you can do the same for cancerous and non-cancerous cells, or for new drug molecules. Automating these processes could help us all.

Competitive advantage

I believe AI and ML will be a major competitive advantage for companies moving forward. And the benefits will not be restricted to one specific industry. Let me give you a few examples from among our customers.

A major credit card company in the US is bringing ML to its various business units. They are creating ML models to predict and control fraudulent credit card applications, accelerate credit card limits to their low-risk customers through automation and optimise marketing spend by processing and analysing large amounts of digital data.

A handful of life sciences companies are leveraging our platform to accelerate early-stage drug discovery by combining genetic sequencing data with (anonymised) disease data of patients to find patterns and opportunities for highly targeted drugs.

Gaming companies, media and retailers are using data across digital and brick-and-mortar channels to optimise customer journeys and present the right offers to the right targets to increase conversion rates and top line.

Leading the charge

The banking and financial services sector is always looking at how it can make use of new technologies. The banks have a lot of data, and they are constantly looking at how they can keep ahead of their competition - AI is one pathway to new service creation and also reducing costs. With all the new regulations coming in around access to data by customers, banks have to look for ways to use this data themselves more effectively while also respecting data privacy.

The healthcare and life science sector is investing pretty heavily in AI and ML technologies as this can help them in their pattern matching, leading to discoveries faster. Manufacturing and industrial companies are looking at ML as part of their "Industry 4.0" strategies.

Companies with lots of customer, user or device data are generally at the top of the list. So this tends to span several industries including retailers, media and entertainment firms, and Adtech or marketing data firms.

Maybe the most surprising is how ready the public sector is for AI. Government organisations can use AI in interesting ways - for healthcare provisioning, security and better operational efficiencies.

Unicorn hunt

I believe that AI has a "one per cent" problem. Until now, only a handful of companies have been able to realise the full potential of AI - companies like Amazon, Google, Facebook, Netflix and Uber - because only they have easy access to a vast amount of cloud data and an army of data scientists.

But that is starting to change with technologies like Spark that make it easier for a range of users to be productive with big data and machine learning models - and to collaborate more effectively.

We need to remember also that AI is a team sport. It requires different skillsets. Companies should not be looking for unicorns, because they don't exist. Successful companies bring together a diverse team of experts: data engineers who curate large data sets and run the complex data infrastructure; data scientists who typically have a strong math background and look for patterns in the data; and finally domain experts who have deep business expertise and help frame the problems that need to be solved.

The key is to empower these teams to collaborate, easily get access to the data, and quickly iterate to find the right models that will power the business.

London: AI central

I think London is quite possibly the best European location to support AI initiatives. At Databricks we will be working with organisations all across EMEA on their AI initiatives, but very intentionally chose London for our new EMEA head office. There are companies in the technology sector like DeepMind here already.

There are also head offices for creative and advertising companies that have strong data and analytics teams in place, and we have multiple universities that support new graduates in these exact efforts. The open source community is strong and continues to develop further - and there is the pull of an international city.

For Brexit, there are some uncertainties in place, both for companies and individuals. However, we chose London despite Brexit because its strength for the business community and its ability to be a hub for operations is simply unmatched.

The positives around the economy, the innovative minds and businesses here, and the market opportunities all mean that this is an ideal environment for growth and ultimately, for business success.

Cutting through the hype

There's a lot of hype to wade through. Tech teams that stay focused on how to achieve AI, and how to be productive with ML or Deep Learning often have a helpful perspective to cut through the hype.

We see a lot of companies creating smaller innovation centres to experiment and test the waters with an initial project that has an immediate business impact. As soon as they have success, they are expanding the capability and exploring more use cases.

The common thread we are seeing is that companies believe in the promise but they are struggling with the "how".

David Wyatt is general manager EMEA at Databricks