Use Cases

Working with innovative clients to bring AI to light

The Roadmap to AI Enlightenment

Artificial intelligence, and data science in general, are experimental in nature. The number one reason we see AI projects fail is that they’re treated as black boxes and organizations don’t take the time to experiment with what works and what doesn’t.

Before you run with AI, you must learn to walk.  AI Discovery, our signature process, helps organizations derive value from their data more quickly by laying a strong foundation on which to build up to AI success.

What is the “data science process?”

People are at the core of every business, making decisions and taking actions that impact the bottom line. Data Science is the discipline of identifying opportunities within business processes and using data and technology to facilitate more profitable decisions. It complements and extends the expertise of individuals to create sustainable value within an organization. Our process:

  • Addresses specific business challenges

  • Transforms data to meaningful information

  • Informs decisions that impact the bottom line

What do all successful AI projects share?

  • Buy-in. The organization is aligned on ROI and AI projects are prioritized according to organizational goals.

  • Trust. Solutions are delivered with appropriate amounts of education and stakeholders understand the underlying assumptions.

  • Consistent Performance. Performance and value creation are consistently tested and different users of the solution experience similar results.

What do all successful AI projects share?

  • Buy-in. The organization is aligned on ROI and AI projects are prioritized according to organizational goals.

  • Trust. Solutions are delivered with appropriate amounts of education and stakeholders understand the underlying assumptions.

  • Consistent Performance. Performance and value creation are consistently tested and different users of the solution experience similar results.

So, where do I start?

Ask one simple question…

If I could identify  ,
I could  

(ex. If I could identify when a customer becomes a higher risk for churn, I could alert our retention team to intervene.)

So, where do I start?

Ask one simple question…

If I could identify  ,
I could  

(ex. If I could identify when a customer becomes a higher risk for churn, I could alert our retention team to intervene.)

How about some successful AI use cases?

Using AI to improve the customer experience

When a fortune-500 company decided to double down on customer experience and collect consistent feedback more often, their customers gladly shared their opinions. Segmenting existing customers into promoters and detractors is powerful. Studies have shown that promoters are 6 times as likely to buy, while detractors are responsible for more than 80% of negative word of mouth. They went from a few hundred satisfaction survey responses a month to over 4000, a quarter of which were in a foreign language. While the satisfaction metrics were informative, the true value was in the qualitative feedback. Their customers were telling them exactly what they were enthusiastic about…and what they weren’t. However, without a quantitative way of looking at these responses, the root causes were difficult to spot.

Pandata used artificial intelligence to build a solution that identified critical customer experience issues in qualitative feedback, highlighting the most relevant positive and negative comments. Whereas divisions previously viewed data shared with them as anecdotal, the organization now has regular conversations around these qualitative responses. They are quicker to respond to customer needs, and as a result, their Net Promoter Score has increased by 20%.

Using AI to detect insider threat

The typical cybersecurity application involves safeguards and mechanisms to protect against outside compromise. However, insider threat can be extremely destructive – when an employee is compromised and is acting from within or an employee is acting in a non-malicious yet risky manner that leaves the organization vulnerable to attack. Insider threat can be difficult to detect as a compromised employee is acting permissibly and is therefore not explicitly breaking rules yet is creating a pattern of suspicious behavior.

Pandata partnered with FirstEnergy (a Fortune 500 utility company) to develop an AI solution around insider threat detection to create a Holistic Risk Profile for all employees based on physical and digital behavior. This approach builds on user behavior analytics – profiling behavior to determine first what constitutes normal, then determine what is abnormal, and finally, what is malicious.

Through partnership with the internal cybersecurity analysts, we incorporated human expertise alongside machine learning to develop a model that both detects abnormality and attributes risk to patterns of behavior. While the work is still ongoing, this AI solution reduces the number of events a cybersecurity analyst needs to investigate from tens of thousands to tens, approximately half of which are noteworthy and would have otherwise been missed.

Using AI & curated content to qualify purchasing decisions

When customers visit your website, they are telling a story about their intent. Marketers have long offered content that requires an email address to qualify prospects and customers. But what if you could qualify more complex behaviors like product or service considerations and readiness to buy? One fortune-500 B2B company decided to do exactly that by transforming the website analytics data into a powerful tool to identify intent.

With more than 500 million website interactions across 3.5 million unique website visitors a year, the data was too complex for traditional analysis. Pandata paired their website data with data from their CRM and used artificial intelligence to identify digital activity that resulted in specific customer actions like contacting customer service, making a purchase, and contacting technical support. The e-commerce team is now able to proactively identify anonymous website visitors demonstrating an intent to purchase within 30 days and with over 80% accuracy. This provides opportunities to target these customers with meaningful content, capturing their contact information and better qualifying them for
sales.

Using AI & transactional data to target and qualify customers

A benefits management company partners with insurance companies to enroll members in special government funded programs. But to do this, they must analyze billions of medical records to pre-qualify which members to reach out to. An analogy in the retail space would be analyzing historical transaction data to target customers for special offers.

With over 200,000 unique types of member data (think 200,000 columns in a database), this data is highly complex, and like most transactional data, imperfect and incomplete. Pandata used artificial intelligence to mine the transactional records of all members that had been either successfully enrolled or denied in the past. The resulting model is helping them uncover an additional 30%+ of qualifying members and significantly reducing the amount of time spent on members that are not likely to be approved.

Cleveland Museum of Art – Using AI to talk about art over time

The Cleveland Museum of Art is a world-renowned art museum with a substantial collection of over 61,000 artworks. In January of 2019, they launched their Open Access Initiative in which they made over 30,000 public domain works and metadata for their entire collection public and downloadable (GitHub). Metadata includes title, description, artist, year, and department, among other details. As part of the Open Access launch, CMA asked Pandata to participate in demonstrating the power of such a data set. Using natural language processing and data visualization techniques, Pandata used the text descriptions from all art that had one (approximately 10,000 works) to visualize how we write about art across time and cultures.  Learn more

Many teams are eager to get started with artificial intelligence projects—often struggling to see the return. If this is you, don’t give up! AI start with experimentation, which thrives on trial and error.
Let Pandata take frustration out of the equation and help you get on the road to AI success. Get started now!