Illustrative engagement across five service pillars and five industries, reflecting representative Meruksha engagement patterns.
Industry: Specialty Retail / CPG
Client Profile
The client is a specialty retail and consumer packaged goods brand generating approximately $500 million in annual revenue, operating a hybrid cloud environment with meaningful workloads on both AWS and Microsoft Azure across its e-commerce and corporate technology functions.
Challenge
Over roughly a year, two separate business functions had independently stood up enterprise AI platforms without any central coordination. The marketing organization had adopted Azure OpenAI Service to power content generation for product descriptions and campaign copy, while the e-commerce technology team had separately adopted AWS Bedrock to drive product recommendation models on the company’s online storefront. Neither team had visibility into the other’s platform usage, cost trajectory, or access control configuration, and corporate IT had no unified identity governance spanning the two environments — each platform had its own access provisioning process, managed by different teams with different offboarding procedures. This created both a straightforward cost problem, since each team was independently building and maintaining data pipelines to feed their respective AI platforms from overlapping source data, and a more serious security gap: because access provisioning and deprovisioning were not centrally managed, there was no reliable enterprise-wide mechanism ensuring that departing employees’ or contractors’ access to either AI platform was revoked promptly.
Approach
Rather than forcing a single-cloud consolidation that would have required marketing or e-commerce to rebuild functioning systems on a platform not suited to their use case, Meruksha designed a multi-cloud AI platform strategy that kept both AWS Bedrock and Azure OpenAI Service in place — each was, in fact, the stronger fit for its respective workload — while unifying identity and access governance across both using the company’s existing enterprise IAM investment rather than standing up new access infrastructure. Meruksha also consolidated the two teams’ separate, overlapping data extraction pipelines into a single governed data lake that fed both AI platforms from one source of truth, eliminating duplicated extract-transform-load work. As part of the engagement, a full access audit was conducted across both platforms as a baseline before the new governance model went live.
Outcome
That baseline audit surfaced that a contractor who had left the company several months earlier still had active, unrevoked credentials on one of the two AI platforms — a gap that was closed immediately as part of the engagement. Beyond the immediate security fix, the consolidated data pipeline eliminated an estimated $180,000 in annual duplicated data infrastructure costs. Marketing and e-commerce continue to operate on the AI platforms best suited to their respective needs, but access provisioning, deprovisioning, and cost visibility for both now run through a single, centrally governed identity and reporting layer.