Flick through the pages of Wired on any given Tuesday and you’ll read that master debating robots, invisible drones and disease-detecting fish are ushering in life 3.0.
But beneath the sensational latest from the likes of Google, Tesla, Amazon or IBM, there are huge business opportunities in mainstream AI. From start-ups to SMEs, firms big and small have the tools of big data and AI at their fingertips – and the bar to entry has never been lower.
Amazon Web Services, Microsoft’s Distributed Machine Learning Toolkit, Google Cloud’s AutoML, Kaggle and IBM’s Watson Developer Cloud – there’s now a plethora of open-access platforms enabling humbler companies to start rolling data innovation into their future.
With accessible costs and tools, almost any company can experiment, build and deploy AI and its related tech. But key to starting that journey is organising and readying your data.
AI is all around
The perceived definition of AI (robots and self-driving vehicles) is different from the reality. There are actually three key parts to AI: perception (where a machine collects data about its environment), cognition (where a machine processes data it collects), and action (where a machine affects its environment in some way).
When all three components work in harmony, you get what’s called ‘Super’ AI. This is the humanoid, ‘thinking’ robot bit. It’s also the bit that scares people, including Stephen Hawking and Elon Musk who have both called for safeguards to prevent a robo apocalypse.
But the majority of AI units comprise just one or maybe two portions of perception, cognition and action. Deployment of these forms of AI is mundane by comparison, yet these breakthroughs are capable of driving business efficiencies and opportunities for the many, not just the few.
Perception: AI and business
Image and object detection are examples of perception AI. Techniques like these help publishers and brands save masses of time and effort by automatically batching and tagging media, people and products; whether it’s to organise them or to flag copyright infringement on channels such as YouTube or Instagram.
Perception also covers speech recognition, and Natural Language Processing (NLP) is an interesting concept on its own. NLP allows companies to gauge the true meaning and sentiment underpinning customer or stakeholder words, speech and text.
We’ve probably all experienced a chatbot or voice-activated menus on the telephone – NLP tools are becoming a staple of customer-facing AI.
Cognition: AI and business
Cognition includes the ‘machine learning’ portion of AI, where machines accumulate data, trends and patterns beyond their basic programming. They learn to ‘think’ for themselves. They fold additional factors and contextual variables into their decision-making in order to calculate/ predict outcomes with accuracy.
Banks have long used cognition tools to identify fraud. Insurers too would be lost without their risk and decisioning dashboards. Ditto City workers and investors. Programmatic advertising is a mainstream iteration of cognition AI as it, through multiple data points, digitally marries products and customers that seem like a good fit based on data, real time and legacy.
Healthcare technologies are progressing at pace based on ‘thinking’ AI. We’re now in an era of 24/7 patient monitoring where patient criteria and trends are fed straight through to the professionals. Plugging the Internet of Things into cognition technologies can let firms glean feedback, patterns and further value from a range of FMCG and household items. The resulting data can be used across the business so companies can learn more about their customers.
Action: AI in business
This is the robotics part. It’s not new in itself – we’ve had mechanised production lines for generations – but when basic robotic technology is married with Artificial Intelligence, units can do much more than one repetitive task: they can multitask and even adapt to their environment. This dynamic continues to drive efficiencies in big retail and manufacturing.
Action-based AI manifests in many ways. And while drones and autonomous vehicles grab headlines, simpler robots now routinely help in a range of sectors. In advanced hospitals, travelling robots enable doctors to consult patients remotely, while software bots are common in corporations and in government, saving costs and man hours by completing repetitive back office tasks.
Several press organisations are now using AI software to write news-in-briefs. Although unequipped to scribe in-depth articles, machines’ ability to write factual summaries and reports based on limited data could appeal to companies of all stripes.
What is data warehousing?
The opportunities in AI are accessible and plentiful. And real world examples of humbler AI are all around. But in order to take advantage of the low bar-to-entry of AI and its subsets – now or in future – companies must set the right data foundations.
There’s increasing focus on the data warehouse as the starting point to innovation in this field. By gathering and inventorying the structured data from multiple sources all across an organisation, firms can house their valuable, lucrative data and use it as the basis for analysis and strategy.
In a sense, the data warehouse is the terminus for all the approved data flowing within an organisation. It is the platform where meaningful analysis of performance, trends and correlations can be drawn to produce razor sharp business insights. The data warehouse highlights long term trends so firms have a bedrock on which to expose gaps and opportunities that can be bridged and met via new ideas.
AI in your organisation
Earlier in this piece we cited a handful of big companies who regularly turn heads with their AI innovations. But the truth is that these companies – with their bombastic, headline-grabbing breakthroughs – are the ones driving and democratising the AI movement.
Not only have they gifted tools of the trade to the rest of us, they stewarding innovation by shining a light on the pinnacle of what’s possible.
For those who think the opportunity to hop on board the AI train has gone, think again. Most organisations who are investigating big data’s disciplines are still in the early stages. There’s time, tools and expertise all around for any business to fold AI thinking into their future provided their data foundations are strong and stable. In many ways, that’s the tough part.
Please get in touch if you want to add order to your data before investigating a future built, at least a little, on artificial intelligence.