SAS Innovate 2025 ran May 6–9 in Orlando, Florida. Its central message was that enterprise AI matters most when it improves real decisions—with data, rules, oversight and governance around it—not simply when it generates content. This is a historical recap of the event, not a live updates page.
What was SAS Innovate 2025?
SAS’s flagship data-and-AI conference brought together technical users, business leaders, customers and partners. Before the event, SAS said it expected more than 3,000 attendees and planned more than 200 breakout sessions, workshops and related activities; those were announced expectations, not independently audited final totals. The program included mainstage presentations, industry sessions and new “Solution Connects” focused on Risk & Fraud, Health Care & Life Sciences, IoT and Customer Intelligence. SAS’s event announcement also listed organizations including Truist, Georgia-Pacific, Norwegian Cruise Line Holdings, Lockheed Martin, Epic Games, Liberty Mutual, Macy’s, Procter & Gamble and Wells Fargo.
The advertised speakers included SAS co-founder and CEO Jim Goodnight, Microsoft chairman and CEO Satya Nadella, Brené Brown, Frank Abagnale, Alfonso Ribeiro and DJ Jazzy Jeff. SAS described the Nadella–Goodnight session as special and prerecorded, so it should not be mistaken for a live in-person keynote from the Orlando stage.
The keynote: AI as decision intelligence
SAS executive Bryan Harris argued that AI’s business value lies in decision intelligence: combining data, analytics, AI and business rules to make better operational decisions. The framing was broader than generative AI. A generated answer may be useful, but in a business workflow the organization also needs to know what data informed it, whether it fits the context, how a decision was reached and who is accountable for acting on it.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
That distinction matters because a technically capable model can still produce poor or biased results when its inputs are inappropriate or biased. In its account of the keynote, SAS emphasized outcomes, explainability and governance alongside AI capability.
Agentic AI: from assistance to autonomous action
At the conference, “agentic AI” described a spectrum: systems that assist a person and keep them in the loop at one end, and systems that take more autonomous action at the other. Discussed possibilities included fraud decisions, risk scoring, customer interactions and operational workflows. The practical proposition was not an unconstrained chatbot, but AI working with models, rules, APIs, monitoring and governance.
The live coverage showed demonstrations involving mortgage cases, decision explanations, model cards and decision lineage. These examples illustrated the direction of travel; they do not establish that SAS had released one generally available autonomous-agent product covering all those scenarios. SAS’s Intelligent Decisioning page describes a product approach that combines AI, machine learning and business rules for operational decisions, but that positioning should not be confused with a specific new launch at the conference.
Product and customer demonstrations
SAS Viya Workbench and text analysis
A stage demonstration used product-review data for sentiment analysis, including stemming and lemmatization. The presenter moved between Python and R and showed the workflow in SAS Viya Workbench, with reference to using a more advanced language model for sentiment scoring. This was a demonstration, not independent evidence that the environment makes development universally faster or easier. Teams evaluating it should check current availability, supported languages, deployment options and licensing for their circumstances. SAS describes the broader Viya platform separately.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Procter & Gamble: a hybrid optimization example
A SAS and Procter & Gamble conference example concerned product-formulation constraints. As reported in the event coverage, a traditional solver took roughly six hours, a quantum-AI approach about two minutes but produced unwanted results, and a hybrid method—using quantum methods for much of the process and a traditional solver for final calculations—took about 12 minutes. The example’s point was the trade-off: the fastest result was not necessarily useful, and a hybrid workflow could balance speed and solution quality. These figures describe that presented problem, not a general quantum-computing benchmark.
Georgia-Pacific and Epic Games: a digital twin
SAS described a digital twin of automated guided vehicles (AGVs) at Georgia-Pacific’s Savannah River Mill, built with SAS Viya and Unreal Engine. The simulation let managers test layouts and routes, adjust the vehicle fleet and switch between real and synthetic data. In the modeled scenario, the reported optimum was 47 AGVs, associated with an 8% performance improvement. Those are results for that model, not a general manufacturing rule. A digital twin is only as useful as its operational data, assumptions and simulation model.
Orlando Magic: customer engagement
The Orlando Magic example connected targeted fan emails with ticket-resale operations. Season-ticket holders could offer unused tickets in exchange for virtual credit, according to the live coverage. It showed how analytics and customer communications can support a business process as well as outreach; it did not establish a measured revenue result.
Governance, shadow AI and healthcare
Managing AI use at work
SAS positioned trust, transparency and governance as prerequisites for enterprise adoption. Event coverage reported a claim that 58% of employees were already using AI for work, and that 60% of those users relied on tools their employers had not approved. Without the underlying study’s geography, sample and methodology established here, those figures should be treated as claims made during the event—not universal or current statistics.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Unapproved tools can create risks such as confidential-data exposure, inconsistent outputs, missing audit trails and regulatory problems. Governance therefore needs to cover more than a model: it can include data permissions and lineage, prompts, agents, versions, monitoring, decision records and human accountability.
REAHL and responsible healthcare AI
SAS, Erasmus University Medical Center and Delft University of Technology were involved in the Responsible and Ethical AI in Healthcare Lab (REAHL). The initiative’s stated aims included transparency and tracking which models are deployed and for what intended purposes. Topics discussed included drug-safety analysis, medical simulations and cost-of-care assessment.
One controlled test described in the live coverage showed how a hospital IT update could change the data feeding a model and affect its reliability. It was not a reported patient-harm incident. Its broader operational lesson is that a model’s approval is not permanent: changes to software, data, clinical practice or patient populations may affect whether it remains safe and useful. Model registries and ongoing monitoring help make those changes visible.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the examples show—and what they do not
| Customer or partner | Area | Reported example | What to take from it |
|---|---|---|---|
| Procter & Gamble | Manufacturing and optimization | Hybrid quantum and traditional solving for formulation constraints | The fastest method may not deliver an acceptable answer; solution quality matters alongside runtime. |
| Georgia-Pacific and Epic Games | Manufacturing and digital twins | Simulation of AGV fleet size and routing | Simulation can test operational changes before deployment, subject to model quality. |
| Orlando Magic | Sports and customer engagement | Targeted fan emails and a workflow for unused tickets | Analytics can link customer outreach with an operating process. |
| Erasmus University Medical Center and Delft University of Technology | Healthcare | Responsible AI initiative focused on transparency and model tracking | Clinical AI needs lifecycle oversight, not just an initial review. |
| Truist, Wells Fargo and other listed organizations | Financial services | Organizations featured in the event’s customer and industry program | Risk, fraud and decisioning were prominent themes; the listing alone does not establish a particular deployment or result. |
The examples are useful for understanding the problems SAS wanted to address, but a keynote or stage demo is not a production benchmark. The event material covered strategy, demonstrations and customer stories; it does not establish that every capability shown was generally available or that the reported outcomes will transfer to another organization.
Best Value
How to assess the technology for your organization
The right evaluation depends on the problem, the operating environment and the controls required. Before considering a platform or agent, ask:
- Problem fit: Is the need prediction, optimization, decision automation, simulation, governance or content generation?
- Data readiness: Are data quality, permissions, lineage and refresh frequency adequate for the decision?
- Oversight: Which actions can be automated, and where is human review required?
- Explainability and audit: Can the organization explain why a model or agent produced a result and reconstruct the decision later?
- Deployment and integration: Does the solution fit public cloud, private cloud, hybrid or on-premises requirements, and connect to existing data and workflow systems?
- Skills and portability: Does the team work in SAS, Python, R, SQL or low-code tools, and how much dependence on SAS-specific workflows is acceptable?
- Regulatory burden: Lending, fraud, insurance, healthcare and public-sector use cases typically demand especially strong documentation and controls.
- Total cost: Account for licensing, infrastructure, implementation, migration, training, monitoring and support—not software price alone.
SAS’s integrated platform approach may appeal to organizations that value managed analytics, governance and decisioning in one ecosystem. A modular or open-source stack may offer more developer choice, but can require the organization to assemble and operate more of the surrounding workflow. Cloud convenience can also conflict with data-residency or infrastructure requirements, while more autonomous systems can reduce manual work but make errors more consequential.
What was actually new?
The event mixed product demonstrations, customer examples and strategic direction. The clearest distinction for readers is between what was shown and what was established as a product release: Viya Workbench was demonstrated; the quantum and digital-twin figures were scenario-specific customer examples; agentic AI was a broad capability theme; and Intelligent Decisioning is an existing product positioning, not evidence of a new launch at Innovate 2025. The event material cited here does not establish general availability, licensing or deployment details for every demonstrated capability.
For buyers, the useful next step is to compare deployment options, governance needs and total cost against the systems already in place. SAS’s product pages provide routes to explore Viya and Intelligent Decisioning, but a conference demonstration by itself cannot establish product fit or expected results.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
