Reduces assessment prep time from up to six hours to one with Gemini
Contributes to a 20-point increase in its customer Net Promoter Score
Delivered a functional prototype of the custom AI agent in one week
Fine-tuned a custom model with thousands of records within four hours with Google Cloud
Cognizant is embedding its institutional knowledge into AI to empower auditors and enhance critical engineering quality assurance, ultimately delivering greater value and satisfaction to clients.
Cognizant is embedding its institutional knowledge into AI to empower auditors and enhance critical engineering quality assurance, ultimately delivering greater value and satisfaction to clients.
The new way to ensure consistent quality audits across thousands of projects
How can a global enterprise ensure consistent quality and efficiency across 18,000 active projects? Expert auditors at global technology consultancy Cognizant were spending up to six hours on manual preparation for each assessment, a major drain on efficiency. To ensure quality and empower its team, the company built AI assistants on Vertex AI, fine-tuning Gemini models with its own institutional knowledge. Now, auditors can instantly get help validating evidence, allowing them to focus on analysis rather than data gathering.
For an enterprise, there are two critical aspects to using LLMs: you need accurate capability when handling large datasets, and you need consistent responses. We found that Google Cloud was able to give a better solution on both.
Pandiyan Adiyapatham
Business CIO for Delivery Excellence, Cognizant
Why Google Cloud?
Out of multiple solutions considered, Cognizant chose Google Cloud for its superior accuracy, consistency, and ease of use. In particular, what stood out was the ability to easily fine-tune Gemini models on its own proprietary data and to teach AI to "speak the same language" as Cognizant experts. Relying on thorough documentation and robust SDKs and libraries, the team created a successful Proof of Concept in one week, fast-tracking the production of the final solution.
From a technical perspective, the ease of use of Google solutions is amazing. The SDKs and libraries make life very simple, and we were able to solve issues quickly and move ahead with our project.
Krishnamoorthy R
Solution Architect, Cognizant
Building enterprise trust in generative AI
For Cognizant, governance and output reliability were key factors in their generative AI adoption journey. To address these concerns and align with domain-specific expectations, the team implemented robust validation workflows and a series of checkpoints to measure response quality. On the workforce front, they equipped their teams with targeted upskilling programs in ethical AI practices and model interpretability to maintain high standards of trust and deliver on client expectations.
We undergo reviews to ensure that the solution we're looking to implement is following all the responsible AI principles. We also have the human in the loop and a feedback mechanism enabled through this whole process.
Rajasekaran M
IT Head of Delivery Audit and Assessment Functions, Cognizant
