Amazon Web Services unveiled two agentic AI tools on Monday that package the e-commerce giant’s hard-won operational know-how into products other companies can buy. Connect Decisions handles supply chain disruptions. Connect Talent runs voice interviews for high-volume recruiting. Both draw directly from Amazon’s internal systems, honed on managing 400 million products and onboarding 250,000 seasonal workers last peak season. GeekWire broke the news from AWS’s San Francisco event, where executives framed the launches as a pivot from cloud infrastructure to full business applications.
Colleen Aubrey, AWS senior vice president of applied AI solutions, called it a ‘day zero’ for the applications team. After two years building out the group, she said, ‘If we’re lucky, we’ll have some hits in this collection of four.’ That quartet includes Amazon Connect, now a billion-dollar contact center platform since its 2017 debut; Connect Health, announced last month; and these newcomers. Aubrey likened the strategy to Amazon’s retail marketplace, where house brands sit alongside third-party goods, or Prime Video’s mix of originals and licensed shows. It’s a hybrid model. AWS sells its own apps even as customers build on its foundational services like Bedrock.
Take Connect Decisions. It deploys more than 25 specialized models, including a foundation model from Amazon’s Supply Chain Optimization Technologies team, known as SCOT. The system spots issues like supplier delays or demand surges, ranks them by urgency, and proposes fixes complete with cost estimates and trade-offs. One early user already runs what-if scenarios in business meetings. Their procurement team wants more. AWS plans to deliver.
Connect Talent targets manufacturers, logistics firms, retailers, and hotels chasing entry-level hires. AI agents field voice interviews around the clock. No scheduling hassles. They strip names and resumes for bias reduction, then hand recruiters competency scores and full transcripts. Amazon built this from its own seasonal hiring blitzes. Scale matters here.
But why buy these when you could assemble something on Bedrock? Aubrey answered directly: ‘The complexity of transforming an entire business function, not just an individual task, calls for a purpose-built product that can be used across an organization.’ Point taken. Agentic AI demands orchestration at enterprise scale. Raw models fall short.
The timing fits AWS’s broader push. CEO Matt Garman used the event to spotlight OpenAI models landing on Bedrock for the first time, thanks to a revamped Microsoft-OpenAI pact and Amazon’s up-to-$50 billion investment. Separately, Amazon Q got a desktop app, custom dashboards, and ties to Google Workspace, Microsoft 365, Zoom, and Salesforce. These moves position AWS against enterprise giants—some of them customers. Risky. Familiar territory for Amazon, though.
Amazon’s supply chain AI roots run deep. SCOT’s models power internal forecasting across hundreds of millions of SKUs daily, as noted in earlier AboutAmazon updates from 2025. AWS Supply Chain, a cloud app, already offers ML insights and now generative AI for queries and scenario visualization (AWS). The new Connect tools build on that lineage, productizing Amazon’s edge for outsiders.
Hiring automation echoes Amazon’s past stumbles and triumphs. Back in 2018, a biased recruiting tool scrapped resumes with ‘women’s.’ Now, anonymization and voice scoring aim to fix that. Connect Talent skips corporate executive searches. It’s for volume roles where speed trumps nuance. Industries like hospitality can’t wait weeks per candidate.
Competition heats up. Infor and AWS just rolled industry agents for manufacturing, tackling ERP-tied workflows in scheduling and procurement, per recent X posts from @iblai_. Agentic systems proliferate. Anthropic’s managed agents promise production-scale deployment without engineering armies. Broader trends show AI reshaping labor: Amazon cut 16,000 corporate jobs in January amid efficiency drives, with supply chain optimization teams hit (Reuters).
So AWS enters the fray. Its apps compete head-on with software incumbents. Success isn’t guaranteed—Aubrey admits as much. Yet Amazon’s operational data moat gives these tools credibility. Early adopters test in real crises: delayed shipments, applicant floods. If they deliver, expect copycats. And expansion. Procurement what-ifs today. Factory floors tomorrow.
Agentic AI tests the hybrid model. AWS provides the pipes—Bedrock, Trainium chips. Now it sells the water too. Customers build custom agents; AWS offers turnkey ones. Choice appeals. But lock-in lurks. Amazon knows the game. It wrote the rules for marketplaces.
WebProNews is an iEntry Publication