A leading public polytechnic in the United States—celebrated for engineering, computing, architecture, and design—runs a complex academic and administrative machine on data scattered across admissions, alumni, HR, and research. Those datasets sat in 16+ Snowflake stores and fragmented Power BI dashboards, so leaders had no seamless, real-time way to turn questions into answers. That delay wasn’t abstract—it stalled faculty planning, slowed grant targeting, and blurred alumni engagement decisions.
Generating even a basic KPI meant raising an IT ticket and waiting 2–3 weeks for a custom dashboard—by then, the moment had passed. Ad‑hoc follow‑ups were impossible, compounding the blind spots. Non‑technical staff wrestled with static reports, while the ~$30/user/month licensing of Microsoft 365 Copilot made campus-wide rollout financially unrealistic. Without a responsive, user-friendly insight engine, timely decisions—and ultimately student experience—took the hit.
Insight had to move from 2–3 weeks to under 2 minutes—so decision-making could match the rhythm of campus life.
Friction Named, Path Found
Once the choke point—speed at scale without runaway licensing—was named, the answer wasn’t another dashboard; it was a conversational layer on a governed rail across trusted data. In short: bring the question to the data, not the user to another tool.
The Move — Build a Copilot, Don’t Buy a Burden
Persistent built a Connected Copilot in Microsoft Copilot Studio on a pay‑per‑query model—cutting projected costs by 70–80% versus per‑user licensing. Inside Microsoft Teams or a modern PowerApps app, users now ask in plain English and get contextual answers in 2–4 seconds, even across 16+ Snowflake sources. At that speed, a query feels less like a request—and more like a reflex.
How It Works — Five Beats from Question to Answer
- Ask: Users type a natural-language question in Teams or PowerApps. (Example: “What is the average faculty tenure in the College of Engineering this year versus last?”)
- Understand: Azure OpenAI plus a vector database maps intent to the right schema.
- Fetch: Azure Functions and Python services query Snowflake in real time; results assemble in seconds.
- Respond: Dual-panel output—textual insight on the left, embedded Power BI visual on the right. Persistent’s close collaboration with Microsoft enables native, governed Power BI embedding.
- Access Anywhere: SSO-enabled and mobile-friendly, it lives inside daily workflows.
Behind the scenes, Azure OpenAI’s private endpoint keeps sensitive admissions, HR, and research data inside the university cloud, while the vector database sharpens semantic understanding. When the path from question to answer is this smooth, adoption doesn’t need a campaign—it spreads on its own.
Business Impact — Six Hard Metrics, One Cultural Shift
- 2–3 weeks → <2 minutes: Insight turnaround accelerated executive and faculty decisions.
- 40% performance improvement: Complex queries return in 2–4 seconds.
- 1,000+ users onboarded: From senior faculty to IT, adoption spans technical and non‑technical roles.
- 70–80% cost savings: Pay‑per‑query Generative AI replaces per‑user licenses.
- Conversational + visual side by side: Trust builds when text and charts land together.
- Secure, compliant, private endpoint: Azure OpenAI keeps sensitive data inside the university’s cloud.
- A data-driven reflex takes hold: Decisions—from admissions to research—now move in real time.
Self-service answers—no IT ticket, no wait—ignited a data‑driven reflex across the university.
Next Horizon — From “What Happened?” to “What If?”
The Connected Copilot is now a mission-critical decision support system and is expanding into grant proposal analysis, departmental benchmarking, and alumni engagement strategy—laying the rail for future Generative AI in student success, academic planning, and institutional advancement. With hindsight on tap, foresight becomes the obvious sequel.
Why Persistent — Speed, Fit, and Future‑Proofing
- Calibrated to academic budgets and user profiles.
- Delivered in under 8 weeks, pilot to full rollout.
- Deep expertise in Copilot Studio, Azure OpenAI, Snowflake, and Power Platform.
- Tight Microsoft collaboration for native Power BI embedding.
- On time, within budget, architected for the long game.
By combining conversational AI with governed data access and rapid delivery, Persistent turned a licensing roadblock into a scalable Generative AI runway.
Ready to Move at the Speed of Your Questions?
Retire slow tickets and runaway licenses. Talk to Persistent. We’ll co‑build your Connected Copilot so “Can we get that metric?” becomes “It’s already on your screen.”
