Consultants spent weeks interviewing staff to map undocumented mainframes. Now, they use Claude to analyze code and generate specs.
A global professional services firm with 400,000 employees offering audit, tax, and consulting services.
Legacy and mainframe modernization engagements required consultants to manually review undocumented systems, taking up to six weeks for the discovery...
“Claude is giving us the ability to re-energize the AI mindset in our people. It's a lot more relatable. People feel that they're getting to results faster that are both higher fidelity and also more in tune with where they're going.”
Global professional services network for audit, tax, and consulting.
Anthropic is a technology company specializing in artificial intelligence and machine learning solutions.
PwC's Legacy code analysis is part of this use case:
Related implementations across industries and use cases
Technical debt in large codebases hindered velocity. Now, AI handles execution-heavy updates while engineers retain merge approval.
Deciphering legacy systems bottlenecked senior reviewers. Now, early-career devs use Claude to navigate code, freeing seniors to build.
Sandboxed Python wrappers bottlenecked parsing. Claude Code migrated the 50k-line library to Go in 20 hours, doubling production speed.
Technical debt in large codebases hindered velocity. Now, AI handles execution-heavy updates while engineers retain merge approval.
Deciphering legacy systems bottlenecked senior reviewers. Now, early-career devs use Claude to navigate code, freeing seniors to build.
Confidential review for 120 partners once took several people multiple days. The CEO ran it alone in ChatGPT Work, in minutes.
With 115K annual applicants, managers read resumes between van breakdowns. Now Maria screens, schedules, and offers—managers just decide.
Tournaments running simultaneously meant an hour of manual checks each. AI agents now run them in minutes, freeing the team to be proactive.
Large AI training jobs meant fighting for preemptible slots or leaving campus. Marlowe gave any lab guaranteed multi-node access on demand.
Consultants spent weeks interviewing staff to map undocumented mainframes. Now, they use Claude to analyze code and generate specs.
A global professional services firm with 400,000 employees offering audit, tax, and consulting services.
Legacy and mainframe modernization engagements required consultants to manually review undocumented systems, taking up to six weeks for the discovery...
“Claude is giving us the ability to re-energize the AI mindset in our people. It's a lot more relatable. People feel that they're getting to results faster that are both higher fidelity and also more in tune with where they're going.”
Global professional services network for audit, tax, and consulting.
Anthropic is a technology company specializing in artificial intelligence and machine learning solutions.
PwC's Legacy code analysis is part of this use case:
Related implementations across industries and use cases
Technical debt in large codebases hindered velocity. Now, AI handles execution-heavy updates while engineers retain merge approval.
Deciphering legacy systems bottlenecked senior reviewers. Now, early-career devs use Claude to navigate code, freeing seniors to build.
Sandboxed Python wrappers bottlenecked parsing. Claude Code migrated the 50k-line library to Go in 20 hours, doubling production speed.
Technical debt in large codebases hindered velocity. Now, AI handles execution-heavy updates while engineers retain merge approval.
Deciphering legacy systems bottlenecked senior reviewers. Now, early-career devs use Claude to navigate code, freeing seniors to build.
Confidential review for 120 partners once took several people multiple days. The CEO ran it alone in ChatGPT Work, in minutes.
With 115K annual applicants, managers read resumes between van breakdowns. Now Maria screens, schedules, and offers—managers just decide.
Tournaments running simultaneously meant an hour of manual checks each. AI agents now run them in minutes, freeing the team to be proactive.
Large AI training jobs meant fighting for preemptible slots or leaving campus. Marlowe gave any lab guaranteed multi-node access on demand.