Generative AI is everywhere in the headlines, but for most organizations, the real challenge is making it work with their own knowledge base.
That’s the focus for SAS and its long-time partner Pinnacle, who are applying retrieval-augmented generation (RAG) to put company data directly in the hands of employees, made possible with the use of SAS® Retrieval Agent Manager (RAM). With SAS RAM, users can quickly build and deploy AI agents that unlock the power of the enterprise knowledge base.
In a recent interview, SAS Vice President of IoT Jason Mann and Pinnacle’s Dave Foster explained how the approach is delivering answers rooted in the information companies already trust.
Most company data is locked away
Roughly 80 percent of organizational knowledge exists in unstructured formats – from user manuals to project reports – scattered across file shares, cloud folders and inboxes.
Foster put it simply: “With the SAS RAG [SAS Retrieval Agent Manager] offering, you’re pointing the language model toward your data, so the answers are relevant to your employees or customers, not someone else’s.”
The approach eliminates the guesswork of searching for the right file while reducing dependence on the open web’s inconsistent information. Considering that over 80% of enterprise data is unstructured, this can be a significant productivity gain.
One search replaces hours of hunting
In high-stakes industries, the time it takes to find the right answer can mean the difference between a quick fix and prolonged downtime.
“You go to one place, it searches those documents for you and it gives you the answer,” Foster said.
By making every document instantly searchable, RAG compresses the “detection to correction” cycle and frees up teams to act on insights instead of chasing them. The result is less wasted time and fewer operational delays.
Trusted partners make AI work
Rolling out AI in an enterprise is about technical integration and trust. Pinnacle’s nearly 30-year relationship with SAS, including four consecutive Partner of the Year awards, gives it the credibility to lead that change.
“We get inside with the customer, identify their problems and walk them through the process,” Foster said. “It’s about being open to their ideas and guiding them through exploration.”
That trust shortens the adoption curve and ensures AI becomes part of the workflow rather than a side experiment.
Plug-and-play still needs expertise
SAS RAM is built for plug-and-play deployment, lowering the barrier for companies to test and adopt RAG. But as Mann and Foster both noted, simplicity in setup doesn’t remove the need for thoughtful configuration.
Every organization has its own processes, data quirks and compliance needs. Matching the technology to those realities is where partners like Pinnacle turn an off-the-shelf tool into a strategic advantage.
