Startup makes superforecasting possible with AI
This article originally published at https://blogs.oracle.com/startup/startup-makes-superforecasting-possible-with-ai on December 1, 2020.
Here’s a mathematical problem: The sum of all the individual country GDPs never equals the global GDP. That means forecasting models are flawed from the start, and it’s impacting global supply chain economics in a big way. Entrepreneur Tony Nash found that unacceptable, so he built an AI platform to help businesses “understand the sum of everything” through a highly automated, globally data-intensive solution with zero human bias.
Complete Intelligence, Nash’s Houston-based startup, uses global market data and artificial intelligence to help organizations to visualize financial data, make predictions, adjust plans in the context of a global economy, all on the fly. The globally-integrated, cloud-based AI platform helps purchasing, supply chain planning, and revenue teams make smarter cost and revenue decisions. It’s a way on how to make better business decisions.
“The machines are learning, and many times that has meant deviating from traditionally held consensus beliefs and causality models,” said Nash. “Causal beliefs don’t hold up most of the time—it’s human bias that is holding them up—our AI data is reducing errors and getting closer to the truth, closer to the promise of superforecasting.”
Massive datasets across 1,400 industry sectors
More than 15 billion data points run through the Complete Intelligence platform daily, making hundreds of millions of calculations. Average business forecasting saas software models use 10-12 sector variables. Complete Intelligence, on the other hand, examines variables across 1,400 industry sectors. The robustness gives businesses insights and control they didn’t have before.
“We’ve seen a big shift in how category managers and planning managers are looking at their supply chains,” said Nash. “Companies are taking a closer look at the concentration of supply chains by every variable. Our platform helps companies easily visualize the outlook for their supply chain costs, and helps them pivot quickly.”
Superforecasting brings a modern mindset to an old industry
Australia-based OZ Minerals, a publicly-traded company, is a modern mining company focused on copper with mines in Australia and Brazil. OZ says their modern mantra is more than technology, it’s also a mindset: test, learn, innovate. They wanted to better navigate and understand the multi-faceted copper market, where the connectivity between miner, smelter, product maker, and consumer is incredibly complex and dynamic. They turned to Complete Intelligence.
“I need a firm understanding of both fiscal and monetary policies and foreign exchange rates to understand how commodity prices might react in the future because a depreciating and/or appreciating currency can impact the trade flows, and often very quickly, which might influence decisions we make,” said Luke McFadyen, Manager of Strategy and Economics at OZ Minerals.
“Our copper concentrate produced in Australia and Brazil may end up being refined locally or overseas. And then it is turned into a metal, which then may be turned into a wire or rod, and then used in an electric vehicle sold in New York, an air conditioner sold in Johannesburg, or used in the motor of a wind turbine in Denmark,” he explains. “The copper market is an incredibly complex system.”
With Complete Intelligence, McFadyen has a new opportunity to test for a bigger-picture understanding and responsiveness. Previously, he updated his models every few months. Now he could do it every 47 minutes if he needed to.
McFadyen points to the impact of COVID-19 as a “Black Swan” event that no business forecasting saas software could have predicted, but is nonetheless impacting currencies, foreign exchanges, and cost curves throughout global copper market and supply chains.
“If your model isn’t dynamic and responsive in events like we are experiencing today, then it is not insightful. If it’s not insightful, it’s not influencing and informing decisions,” he said. “Complete Intelligence provides a different insight compared to how the traditional price and foreign exchange models work.”
McFadyen says early results have reflected reductions in error rates and improved responsiveness.
Cloud power and partnership
Complete Intelligence needed a strong technology partner but also one with global expertise in enterprise sales and marketing that could help boost their business. They found it with Oracle for Startups.
“We have lots of concurrent and parallel processes with very large data volumes,” said Nash. “We are checking historical data against thousands of variables, anomaly detections, massive calculations processing, and storage. And it’s all optimized with Oracle Cloud.”
Nash, who migrated off Google Cloud, says Oracle Cloud gives him the confidence that his solution can handle these workloads and data sets without downtime or performance lapses. The partnership also gives him a credible technology that is native to many clients.
“As we have potential clients that come to us that are using Oracle, having our software on Oracle Cloud infrastructure will make it easier for us to deploy and scale. A seamless client experience is a critical success factor for us.”
Nash says the Oracle startup program‘s free cloud credits and 70% discount has allowed them to save costs while increasing value to customers. He also takes advantage of the program’s resources including introductions to customers and marketing and PR support.
“We’ve been impressed by the resources and dedication of Oracle for Startups team,” he said. “I’d recommend it, especially for AI and data startups ready for global scale.”
Beyond mining: superforecasting futures with AI
Beyond mining, Complete Intelligence is working with customers in oil and gas, chemicals, electronics, food and beverages, and industrial manufacturing. From packaging to polymers and sugar to sensors, these customers use Complete Intelligence for cost and revenue planning, purchasing and supply chain proactive planning, risk management, and auditing teams, as well as general market and economic forecasts.
The error rates for Complete Intelligence forecasts in energy and industrial metals performed 9.4% better than consensus forecasts over the same period, and Complete Intelligence continues to add methods to better account for market shocks and volatility.
OZ Minerals’ McFadyen said, “This is the next step in how economists can work in the future with change leading towards better forecasts, which will inform better decisions.”
Nash and Complete Intelligence are betting on it – and building for the future.