QCC Assistant is an AI platform for SAP® custom ABAP code. It reads your code straight from SAP®, works out what has to change for S/4HANA, writes the smallest fix that does it, and shows that the fix compiles. Nothing is written to SAP® until a named person on your team clicks Approve.
EU-hostedNo inbound firewall portsProduction stays read-onlyEvery action logged
Each one comes with findings from the Simplification Database. VBUK/VBUP are gone. KONV became PRCD_ELEMENTS. MATNR is now 40 characters long. Someone has to open every object, find every affected line, rewrite it, check it still compiles, and put it on a transport.
Senior ABAP developers spend months on fixes that follow known patterns.
The custom-code track is often the biggest line in the conversion budget.
Manual edits under deadline pressure break things, and nobody can say afterwards who changed what, or why.
It reads the source, the includes and the findings, then drafts the fix.
Every proposal passes SAP®'s own syntax check before anyone sees it.
A named person reviews the diff and approves it. Only then is the change written, activated and put on a transport.
A fix opened by a person runs as a live conversation: findings, every agent step, the diff with its compile result, the approval and the apply. Want a different approach? Type an instruction and the agent re-runs the fix.
Every custom object across your systems in one searchable view, with KPIs by package and a page per object.
Reads, locates, matches a proven recipe, drafts the smallest change and proves it compiles.
Fix in ECC, fix after conversion, retire, or needs a human — always with a reason.
Watch every agent step, steer it with an instruction, approve in place.
Everything waiting for a decision in one place — by system, state, or "mine".
One approval card per pattern, on one transport — not hundreds of objects one by one.
Findings by triage, fixes by state, applied vs resolved, and AI cost per fix.
"Fix with QCC" from the assessment; status flows back so both tools match.
A small connector inside your network dials out to QCC. No inbound firewall port — the same pattern as SAP®'s Cloud Connector.
Findings and Simplification Items come in from your QTS assessment or existing ATC results.
The agent works through each object — live in a chat, or in the background for large batches.
A named person reviews the diff, the compile result, the target package and the transport.
Lock, write, activate, record on a transport — then read it back byte for byte.
A new assessment or ATC run confirms the finding is gone. Only then is it resolved.
QCC never says "fixed" when it means "proposed". The wording follows the real state of each fix.
The agent has no write tool at all — it can only propose. Only the server applies a change, and only after a person clicks Approve. Machine accounts can propose, never approve.
A separate check compares what the agent says it did with what its tools actually did. Claims "compiles"? There must be a passed syntax check of exactly that source. No match, the run stops.
Before writing, QCC checks SAP® still holds the version it read. If someone changed the object, it stops. No stray transports, and it never releases one on its own.
S/4HANA systems and ECC 7.40. Where a system's write policy forbids writes, QCC still proves the fix and hands it over as a ready-to-apply file.
A "Fix with QCC" button in the Quatelio Transformation Suite opens the object in QCC, already signed in through your own SSO, with no secrets in the link. QTS sends its assessment in and pulls the fix status back.
Explore QTS →“Nothing writes to SAP® without a named human's click. Dev systems only, transported, fully logged.”
Enforced by SAP® authorizations, not only by our software.
Each project's agent runs in an isolated container that can reach exactly two endpoints: its SAP® system and the AI gateway.
Single sign-on through your identity provider, e.g. Microsoft Entra ID. QCC never stores passwords.
SAP® credentials encrypted at rest with a per-tenant key in a key vault. The agent can never read them.
Every agent step, SAP® call and approval is recorded and exportable.
Data stays in the EU, Frankfurt region.
Statistics are grouped by object, never by developer.
Writes limited to test objects by default, widened per system, on purpose.
| Run fixes | Approve | Admin | Read | |
|---|---|---|---|---|
| Admin Users, systems, write policies, costs, audit | — | — | ✓ | ✓ |
| Developer Run fixes, chat with the agent, propose | ✓ | — | — | ✓ |
| IT lead Approve changes, see progress and costs | — | ✓ | — | ✓ |
| Guest Read-only, limited to chosen packages | — | — | — | ✓ |
of findings handled automatically
average time from finding to approved fix
average AI cost per fix, live on the costs page
of changes approved by a person and on a transport
No. It has no write capability. It proposes a change, and a named person on your team approves it.
No. Writes go to dev or sandbox systems only, on a transport. Production users have no change authorization in SAP® itself.
No. A small connector inside your network dials out to QCC. You open no inbound ports.
The finding is marked needs a human with a stated reason. Nothing is silently skipped.
Yes. QCC reads the code and proves the fix compiles. It writes the fix where the system's write policy allows it, and otherwise gives you a ready-made hand-over file.
QCC reads the source directly from your SAP® system through your connector. AI calls go through a single metered gateway, and every prompt and response is logged for your audit trail.
English and German.
See QCC fix a real ABAP object live, from finding to transport, in 20 minutes.