From Reactive to Proactive: Redesigning How AWS Manages Exclusive Customer Deals

Use Case: Proactive Deal Intelligence

After Uno gave finance managers a single source of truth, GDSP made it proactive. An agentic workflow that monitors 5,000+ customer deals and surfaces revenue risks before anyone has to ask.

The Story

Uno solved the fragmentation problem. But even with everything unified, finance managers were still working reactively. Opening Uno, running queries, analyzing results, one deal at a time. With thousands of active deals per analyst, critical signals were getting buried. The ones that actually needed attention often surfaced too late.

So the question became: what if Uno didn't wait for questions?

Starting with Steve

I built the design around a single persona. Steve, a finance manager composited from contextual research. His day looked like this: arrive, check email, open Uno, manually review flagged deals, build a report for leadership, repeat. Most of his time was spent on deals that were fine. The work that actually mattered kept getting pushed to the end of the day.

Designing Proactive Intelligence

The GDSP workflow flipped Steve's day. Instead of querying the system, the system comes to him. It continuously monitors deal health across his entire portfolio, watching for credit stacking, margin erosion, and unusual patterns. When something needs attention, it surfaces a prioritized alert with context: what changed, why it matters, and what to do about it.

The Hardest Problem: Noise Calibration

An alert system that flags everything is useless. I worked closely with finance stakeholders to figure out where to draw the lines. Which signals warrant interrupting someone's day. Which can batch into a daily digest. Which are just good to know.

The goal was for the system to feel like a sharp colleague tapping your shoulder at the right moment, not an alarm blaring constantly.

That also meant making a deliberate call about what the agent handles on its own versus what it puts in front of a human. Routine pattern detection runs quietly in the background. Anything touching a revenue decision gets surfaced with full context so the manager stays in control. That boundary was the most important design decision in the whole project.

From Analysis to Action

The workflow doesn't just surface problems. It generates AI-written insight reports ready for leadership. What used to take hours of manual prep became a single click. Detection, analysis, and reporting all happen without anyone having to drive it.

This was showcased at AWS re:Invent as a flagship example of agentic AI in enterprise finance.

The Impact

• 99% reduction in time on task for enterprise agreement modeling, from 24 hours down to seconds
• 2,000 hours saved in deal modeling so far this year
• $1M in estimated time savings impact
• Projected $100M+ profit impact through a 1% improvement on average net margins across enterprise agreements
• 4 director-level presentations, surfaced directly to the AWS CFO
• 10,000 hours projected saved across deal modeling, health monitoring, and retrospectives

The Connection

Uno and GDSP are two stages of the same idea. First, unify the data and build trust. Then make the system smart enough to act on it. Together they moved AWS Fintech from tools people use to intelligence that works for people.