How automating AI empowers humans to master business innovation
So far, 2019 has been an exciting year for the evolution and adoption of artificial intelligence (AI) as more businesses are tapping into it as a technology-enabled business strategy. According to PwC 2019 AI Predictions, monetization of AI is projected to address key enterprise performance levers: growing revenue and increasing profits (48 percent of survey respondents), creating better customer experiences (46 percent), improving decision making (40 percent), innovating products (39 percent) and achieving cost savings (38 percent).
Recently, IBM debuted AutoAI, a new capability of Watson Studio that automates critical yet time-consuming parts of the data science lifecycle—including data preparation, feature engineering and hyper parameter optimization. The benefits of AutoAI are many, including:
- Expert data scientists can reduce the time to higher prediction accuracy by automating an AutoAI experiment.
- “Citizen” data scientists and new graduates can get started quickly by watching how the best models are selected on the leaderboard.
- Businesses can see better yields from data science investments.
With the automated AI and ML advancements, you may find yourself wondering: what are the overall impacts to business? How will all of this technological progress impact the ways we run our business and perform our jobs?
Learning from our recent history
In 2011, venture capitalist and innovator Marc Andreesen authored an op-ed, “Why software is eating the world.” He predicted that many industries would be fundamentally disrupted by software—either the incumbents would need to become software companies themselves or they would be outmaneuvered by competition from software-driven companies entering their markets.
Now in 2019, Andreesen’s predictions have not only proven accurate, they have clearly forecast the profound impact of modern software. For Amazon, the bookstore market was merely an aperitif: the company now dominates the retail sector and has changed the ways we shop. Uber and Lyft have devoured the personal transportation industry and, in turn, increased our appetite to use point-to-point ride share. Well-established banks, insurers, healthcare and energy companies have invested in hiring the best software talent in-house, as they feel the pressure of competitors nipping at their heels.
Why did this happen? Business reached a tipping point where making software became easier. Any business could use open-source software and tools to serve millions of users at negligible costs, exploiting a level playing-field. Additionally, a hybrid multi-cloud environment helped startups and enterprises alike start small and scale their services to meet customer demand without heavy upfront infrastructure investments.
AutoAI helps model-driven business be more customer-centric
Going a step further, as Steven A. Cohen’s article “Models will run the world” implies, model-driven businesses are empowered to become more customer-centric, because they can create a virtuous cycle where “models improve products, products get used more, [and] this new data improves the product even more.” Conversely, if businesses use AI unethically or ineptly, customers won’t hesitate to move away from them.
This is why we believe in AI for AI – AI designing AI, AI optimizing AI, and AI governing AI – to help businesses drive innovation and achieve breakthroughs.
If you’re looking for an opportunity to incorporate AutoAI, customer-centric use cases can be a good place to start. There are vast opportunities for businesses to use models to monitor and improve customer experience, through real-time sentiment and tone analysis of customer emails and phone calls. This is why visual recognition and natural language classification are some of the most popular IBM Cloud services integrated with Watson Studio.
The human element
The true opportunities of adopting AI may not be what you expect. Some fear that it will take away jobs. True, many businesses initially become interested in AI as a means to automate processes and reduce costs. However, the more likely outcome is that, just as the rise of software opened up opportunities for a new workforce of developers, the rise of AI will create a new set of roles and an evolution of existing roles: data scientists, data engineers, data stewards, application developers, and line-of-business experts. You can read more about this in my blog “Learning and leading in the era of artificial intelligence and machine learning.”
The collaboration of human intelligence and AI is the future of work. AI alone does not produce desired outcomes unless you define experiment and training scope that reflect the domain knowledge and complexity of business. You need to bring your whole organization together, break down silos of tribal knowledge, and gain a clearer vision. You’re not just wrangling datasets for your AI models to train on. You’re building a common business view that your human decision-makers can benefit from too.
Automation and innovation are two sides of the same coin
Innovation and creativity are essential ingredients for new ventures or strategic pivoting. Automation becomes part of our arsenal to speed the tasks that used to take humans inordinate amounts of time.
Even the most talented data scientists often still need weeks and months to examine and select the right neural networks and other algorithms to reach an acceptable level of model accuracy. With automation, data scientists reduce the time needed to select and monitor models to minutes and hours. This allows them to focus on refining project definition and tackling new AI projects instead of manually checking and modifying code to continuously adjust the models.
The mandate for business is to use automation as a tool to reduce tedious work and rely on AI for the type of jobs machines are better suited to do.