Everyday AI Explained, Part 2: Three Core Capabilities

August 10, 2018 · 3 min read
29658

In our previous post, we discussed the five core tenets of what WorkFusion calls Everyday AI — a practical, scalable, self-service approach to automation for the enterprise. Now, beyond telling you about what Everyday AI is, we’re going to show you exactly what it’s capable of doing, and what benefits its powerful features can bring to your business.

Smarter bots

According to McKinsey, a quarter of machine learning use cases require daily manual refreshes and model updates. A third of them need to be refreshed monthly.

On average, it takes a data science team three weeks to produce a data model. But when the inputs are changing daily, there’s a good chance all that time and effort will be wasted. There’s no guarantee the model which worked for your team last week will work for them today.

AutoML helps you overcome these challenges by making machine learning more manageable. In essence, it’s your team’s virtual data scientist.

With its built-in library of patterns, tailored to a variety of business situations, AutoML can automate pattern recognition. It also includes pre-built features and models, so you don’t have to search high and low for algorithms. Best of all, it integrates machine learning into your existing process workflow, so you won’t have to waste time and resources to code, deploy and connect it yourself.

With the performance enhancements in our latest release, SPA 9.0, AutoML is far more efficient; six times faster, over a third more accurate, and uses far less resource-intensive tasks than its predecessor.

Smarter insights

While it’s important to know what’s under the hood, what matters more is how AutoML’s management and performance benefits can enhance your processes. In other words, the enhancements must be measurable. This is why we’ve amped up analytics in SPA 9.0.

Within the Analytics Dashboard, you and your colleagues can monitor everything that happens in the SPA environment. You can see a high-level overview of all your processes — from day-to-day volume to tracking SLAs, to bot utilization, to determining which processes have the most errors. Or, you can hone in on speed, capacity and the progress of each process on an individual basis.

Bringing bots, people and workflow together on one platform doesn’t just improve automation in theory — it allows you to actually see and measure improvements, so you can keep your finger on the pulse of your business and continually apply those insights to future enhancements.

Smarter governance

In April of 2018, the FDA approved the first-ever, AI-powered medical diagnostic device. This product, which can detect diabetic retinopathy in diabetes patients, is able to provide an instantaneous screening decision without verification by a clinician. If the AI says so, a patient can begin to receive care immediately — no questions asked.

Like self-driving cars and facial recognition, the use of this software in medicine is a big step towards trust in artificial intelligence. But that doesn’t mean AI is ready for total autonomy. After all, would you feel comfortable handing complete control of your company’s billing and payment processes over to a computer program? Probably not. This is why we’ve established robust quality controls to oversee AI-driven RPA. We call it AutoQC.

AutoQC uses AI to add a higher level of resiliency to the traditional maker-checker model. It governs data extraction, classification, ranking, anomaly detection, and other bot output and will inform a human checker of any exceptions or partially fulfilled tasks in real time. You’ll control the level of oversight — check anywhere from 0 to 100% of your bot output, or allow AutoQC to automatically assign a percentage to the checker in order to mathematically guarantee a certain degree of accuracy.

On top of that, AutoQC facilitates continuous learning by turning all faulty bot output into input for AutoML’s retraining feature. Simply put, it discovers why exceptions occur and applies that knowledge to tune machine learning accuracy going forward.

By leveraging the three most intelligent capabilities of SPA 9.0 — machine learning via AutoML, the Analytics Dashboard, and AutoQC’s advanced quality control — your organization can harness the full power of Everyday AI and maximize the value of every data element, bot and employee under your roof.


Also published on Medium.

Drive productivity with Intelligent Automation Cloud
Drive productivity with Intelligent Automation Cloud