SAS launches AI supply chain agent in industry push

SAS has expanded its portfolio of industry-focused AI tools with new agents, models and model pipelines. The update includes the SAS Supply Chain Agent, now in private preview for customers.

The new products are part of SAS's USD $1 billion investment in industry solutions and target sectors including manufacturing, retail, government and financial services.

SAS is placing the supply chain tool at the centre of the release. It is designed for supply and operations planning, a process retailers and manufacturers use to manage inventory and respond to shifts in demand and materials availability.

That work is often managed across departments in spreadsheets over several days and typically runs only once a month because of the effort involved. SAS says its new agent runs continuously, letting users test scenarios, review likely outcomes and track operations on a near real-time basis through a chat-based interface.

Kathy Lange, Research Director at IDC's AI, Data, and Automation Software Practise, said the product tackles a more complex operational task than many existing agent tools. "Current pre-packaged agents tend to tackle basic processes; with Supply Chain Agent, SAS is compressing a very complex process, which could deliver significant value," Lange said. "This offering positions SAS to bring its longstanding supply chain knowledge to a new generation of agentic AI solutions."

Digital twins

Alongside the new supply chain product, SAS highlighted its work using digital twins to model industrial settings. It builds virtual replicas of customer facilities in Epic Games' Unreal Engine so clients can test scenarios and examine operational bottlenecks.

One example involves a medical device sterilisation provider working with SAS to create a digital twin of its facility. The customer believed trays of medical tools were being delayed by a buffer lift used to line them up for cleaning, but the virtual model suggested the lift was acting as a central distribution point instead. SAS says targeted changes then removed the bottleneck and increased production pace.

SAS also outlined how it is using synthetic data and computer vision in workplace safety. SAS Worker Safety lets organisations create simulated video footage from digital twins to train models on events such as falls, machinery accidents and failures to wear protective equipment correctly.

Because the footage is simulated, organisations can generate large numbers of variations in lighting, equipment and worker appearance, while also modelling rare incidents for which no real footage may exist. SAS says the approach avoids using real employees' personal data and allows repeated testing of specific actions before models are deployed across cameras in a facility to issue alerts.

Government use

In the public sector, several US states are using SAS's Payment Integrity for Food Assistance product to help manage Supplemental Nutrition Assistance Program payments. The tool is intended to connect with existing state data, identify patterns linked to overpayments or underpayments, and help case workers focus on the highest-priority cases.

Nevada is among the states using the system. The backdrop is tighter federal oversight of payment error rates, with state budgets facing direct fines if they exceed thresholds for mistakes tied to eligibility calculations, case data or fraud.

For supervisors, the product offers a live dashboard rather than periodic reporting, while investigators receive prioritised leads instead of relying entirely on manual reviews. The aim is to reduce missed benefits for eligible families as well as financial penalties for states.

Fraud focus

Financial services remains another core market for SAS. The company cited research conducted with the Association of Certified Fraud Professionals showing that 75% of anti-fraud professionals are seeing a surge in fraud and scams targeting consumers, while 55% expect deepfake social engineering and generative AI document fraud to increase significantly over the next two years.

Only 7% of those surveyed said their organisations were more than moderately prepared to detect or prevent AI-driven fraud. Against that backdrop, SAS is marketing fraud detection models and agents for banks, insurers and other financial institutions.

SAS Fraud Decisioning for Payments is designed to detect fraud in real time across different transaction types. Its fraud models have been trained on consortium data from major financial institutions covering credit card, debit card, ATM, digital wallet and application fraud, as well as money mule activity.

Manisha Khanna, Global Market Strategy Lead, Applied AI at SAS, said the company is trying to move customers away from fragmented AI projects. "When organisations are left stitching together ad-hoc AI frameworks and experiments, they often fail to achieve the competitive edge they're looking for when they invest in AI," Khanna said. "We're engineering industry accelerators with purpose: to solve defined, real industry problems in highly regulated environments.

"With production-ready agents and models that work on data they already have, our customers across industries can and are achieving extraordinary outcomes."