Article by Peter O’Connor, vice president of sales for Asia Pacific, Snowflake Computing
Artificial intelligence (AI) tools are rapidly growing in capability, and many enterprises are keen to put them to work.
However, very few enterprises are actually making progress or enjoying the value these tools can potentially deliver.
For the enterprises that do succeed, the rewards can be significant. They can obtain insights into market trends and customer requirements that would not have been possible previously.
AI tools can also be used to automate many processing tasks, thus freeing staff to focus on more value-adding activities.
For the many enterprises that have yet to reach this stage, there are a number of specific steps they can take to improve their chance of success. By reviewing these steps and putting them into action, an enterprise can quickly begin to enjoy the benefits that AI technology has to offer.
Traditionally, companies have tended to classify data in one of two ways.
The data is either high-fidelity, high-value data (such as customer records and transaction data), or raw, unstructured data that is either archived or deleted.
For decades this approach made sense, but times have now changed. With the rise of cloud platforms, the cost of storing and processing data has fallen dramatically, which makes deleting it just to save money nonsensical.
So, if your company is still practicing data austerity, the first step in your AI journey is to stop.
Data is your most valuable asset, so keep it all. Better yet, start collecting and storing even more because modern approaches to AI are based on training algorithms with large amounts of data.
Generally speaking, the more data you have, the better your AI tools will become.
When data was expensive to store and process, making data-driven decisions was challenging.
It was often too costly to get the right data in front of the right people at the right time. That’s why many businesses tended to rely on the intuition of senior managers when making important decisions.
Today, technology limitations are quickly fading and it’s now possible to collect fine-grained information about how people are interacting with your business, product or service. It’s also possible to make data available in near real time.
However, solving the technical challenges of data storage and processing alone does not guarantee success. To turn the corner on AI, a company must also get comfortable relying less on the intuition of managers.
Making better decisions with data, not intuition, is the entire point of adopting AI.
With more data informing the decision-making process, it’s time to start setting up experiments.
To achieve the best results, it’s important to create a culture where a broad group of individuals and teams (rather than a small set of senior leaders) has access to the data and is not afraid to try out different ideas and see what works.
Most machine learning and AI approaches involve trying out lots of different algorithms with different metrics and parameters and observing the impact they have on specific problems you're trying to solve.
If a business already has a culture of experimentation, adopting machine learning and AI-based practices is relatively straightforward. Chances are you already have processes in place to collect, analyse and use data effectively.
From there, it’s often easy to find opportunities where AI can augment these human-based processes.
When machine learning and AI algorithms are trained with data, it’s best to start with a large, rich, and structured data set that humans can query to answer questions on their own.
Next, train an algorithm with the same data to achieve the same goal.
This might be answering a particular question or identifying certain patterns. Then, evaluate the algorithm against the known, human-answered dataset to determine how well it performs.
As the algorithm gets better, you can start asking it more advanced questions, with the goal of answering questions human trainers are less capable of answering themselves. The key to ensuring optimal results as algorithms become more sophisticated is “closing the loop.” Ensure input data is always good and constantly evaluate outcomes against the right metrics.
The more decisions are based on data, the more important it becomes to identify exactly what it is that you are optimising for, and defining the metrics being tracked.
Generally speaking, choosing metrics that are a proxy for customer satisfaction is a tried-and-true approach.
When measured properly, customer satisfaction can provide insights across every aspect of a business. For example, measuring renewal rates can not only tell you if customers liked your product but also whether they’re happy with the customer support experience.
For the most accurate reflection of customer satisfaction, optimise across a range of metrics.
While it’s tempting to search and optimise for the one “perfect” metric, tread cautiously because over-optimising for a single metric can easily lead to distortions in product behaviour and create new challenges down the road.
By following these steps, a business will be better placed to extract significant value from its AI tools and initiatives.
Getting things right early will also position the business to gain further benefits as the capability of the tools is extended into the future.