Back office ops · Production

PwC accelerates enterprise-scale GenAI adoption with CrewAI, boosting code-generation accuracy from roughly 10% to 70%+

The problem

PwC consultants needed faster, more-accurate generation of proprietary-language code and lengthy spec documents, but early Gen-AI prototypes produced inconsistent results and offered little transparency into ROI, undermining user trust.

First attempt

PwC initially built its own plug-in framework during its firm-wide Gen-AI transformation, but the early prototypes lacked real-time feedback, produced inconsistent results at around 10% accuracy, and offered no transparency into ROI.

Workflow diagram · grounded in source
1
Consultant requests code or document
trigger
“PwC consultants needed faster, more-accurate generation of proprietary-language code and lengthy spec documents”
2
CrewAI agents execute generation
ai_action
“Crew-powered agents boosted code-generation accuracy from 10% to 70%”
3
Agent monitoring captures task metrics
validation
“Native agent-monitoring integrations gave PwC unprecedented visibility into task durations, tool selection, and human-versus-agent effort—crucial for demonstrating ROI”
4
Granular ROI data delivered
output
“supplied granular data to prove ROI”
Reported outcome

Crew-powered agents boosted code-generation accuracy from 10% to 70%+, slashed turnaround time on complex documents, and supplied granular ROI data, restoring consultant trust and accelerating adoption of agentic solutions across PwC.

Reported metrics
Code generation accuracy — baselineroughly 10%
code generation accuracy — with CrewAI70%+
Document turnaround timeslashed turnaround time
Reported stack
Crew AI
Source
https://www.crewai.com/case-studies/pwc-accelerates-enterprise-scale-genai-adoption-with-crewai
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Crew-powered agents boosted code-generation accuracy from 10% to 70%+, slashed turnaround time on complex documents, and supplied granular ROI data, restoring consultant trust and accelerating adoption of agentic solu…

What tools did this team use?

Crew AI.

What results were reported?

Code generation accuracy — baseline: roughly 10%; code generation accuracy — with CrewAI: 70%+; Document turnaround time: slashed turnaround time (source-reported, not independently verified).

What failed first in this deployment?

PwC initially built its own plug-in framework during its firm-wide Gen-AI transformation, but the early prototypes lacked real-time feedback, produced inconsistent results at around 10% accuracy, and offered no transp…

How is this back office ops AI workflow structured?

Consultant requests code or document → CrewAI agents execute generation → Agent monitoring captures task metrics → Granular ROI data delivered.

Related back office ops cases