Watch a realistic AWS modernization play through, days and review gates included
A GSI quotes 5 years and $14M for the full 38-context portfolio. Cursor compresses the agent-doable work and surfaces every human review gate (architecture, security, FinOps, cutover, code) — one bounded context ships to prod in ~22 calendar days with 4 human review gates, the portfolio in ~18 months.
Run AWS modernization readiness scan
Java EE · WebSphere 8.5 · Oracle 12c · on-prem
1@Stateless2@TransactionAttribute(TransactionAttributeType.REQUIRED)3public class OrdersService {45 @PersistenceContext(unitName = "orders-pu")6 private EntityManager em;78 @Resource(mappedName = "jdbc/OracleDS")9 private DataSource oracleDs;1011 public long reserveInventory(long orderId, String sku, int qty)12 throws NamingException, SQLException {1314 Context ctx = new InitialContext();15 DataSource ds = (DataSource) ctx.lookup(16 "java:comp/env/jdbc/OracleDS");1718 CallableStatement cs = ds.getConnection()19 .prepareCall("{ call SP_RESERVE_INVENTORY(?, ?, ?, ?) }");20 cs.setLong(1, orderId);21 cs.setString(2, sku);22 cs.setInt(3, qty);23 cs.registerOutParameter(4, Types.REF_CURSOR);24 cs.execute();2526 ResultSet rs = (ResultSet) cs.getObject(4);27 if (rs != null && rs.next()) {28 return rs.getLong("RESERVATION_ID");29 }30 return -1L;31 }32}
Candidate bounded context: OrdersService
Target: AWS Lambda · Aurora Serverless v2 · CDK · ECS Fargate
MAP eligibility
Migration Acceleration Program · Modernize
Reads the real legacy files from disk — scan latency is not simulated.
What this unlocks for your book.
The customer's buyer sees modernization. The AE / SA / PDM sees three quota levers inside the current fiscal year.
Pulled-forward managed-service ARR
Every EJB or stored proc Cursor modernizes becomes Lambda + Aurora consumption 42 months sooner. ~$11M ARR per Fortune 1000 account, captured inside the current quota year.
MAP engagements that actually finish
MAP credits only settle when workloads modernize. Cursor keeps the in-house team on the keyboard — no GSI middleman, no 5-year timeline. MAP success rate, not just MAP attach.
Bedrock + Amazon Q pull-through
Every Cursor-extracted context lands on Aurora Serverless v2 and Lambda in AWS SDK v3 — already Bedrock-agent and Q Developer ready. Net-new AI ARR on the same account.
Value capture per modernization engagement
Same monolith. 3.3× faster. Managed-service ARR unlocked by Q2, not Q20.
Metric
GSI baseline
With Cursor
Delta
Per bounded context
~16 weeks
~22 calendar days
(4 human review gates included)
5× faster
Portfolio timeline (38 contexts)
5 years
18 months
(GSI-quoted MAP · parallelized waves)
3.3× faster
Services spend
WebSphere + Oracle + DC
Lambda + Aurora + CW
($8.4M → $2.1M annual)
−75%
Managed-service ARR
Deferred ~5y
Pulled forward ~$11M
(quota upside for AE)
42mo earlier
Bedrock / Q pull-through
Blocked by legacy
Unlocked per context
(every extraction is Q-ready)
compound
Scales to every Java EE / .NET / mainframe monolith in the AWS pipeline. Every context Cursor extracts becomes managed-service ARR.
Why an AWS SA trusts this motion
Cloudscape-shaped guardrails, not cowboy migrations.
IAM least-privilege by default
Emitted CDK stacks scope every action to a specific ARN. No `*` resources. No `iam:*`. Codex audits before the PR opens.
No public data plane
Aurora endpoints stay private. Secrets live in Secrets Manager. VPC endpoints replace NAT for Secrets Manager + RDS Data API.
Human approval gate
Agent opens the PR and moves Jira to In Review. A human merges. The customer team stays on the keyboard.
Well-Architected citations required
Every modernization PR cites at least three of the six pillars (OPS, SEC, REL, PERF, COST, SUS) with pillar question IDs.
One click. AWS Knowledge MCP + Bedrock + Composer + Codex coordinated. CDK-managed Lambda + Aurora + Secrets Manager, IAM least-priv, ready for review.This is what Cursor as the modernization layer looks like.