Amazon Connect expands into a set of agentic AI solutions

Key takeaways

Amazon Connect is expanding from a single product into a set of four agentic AI solutions designed to work within your existing workflows.
The set includes Amazon Connect Decisions (supply chains), Talent (hiring), Customer (customer experience), and Health (health care).
Each solution is built on decades of Amazon's own experience managing complex operations at scale.
The original Amazon Connect, used by companies like State Farm and Air Canada, is now Amazon Connect Customer.

At AWS, we've always believed in putting the most powerful technology in the hands of anyone who needs it. AI has never been more capable or more accessible. But turning that capability into an impactful part of how your team works is hard. Most solutions require you to rebuild your processes around them, demanding extensive change management, technical overhauls, and months before you see results. The result is a gap between powerful technology and the people it was built to help.

We're expanding Amazon Connect into a set of four agentic AI solutions, each purpose-built for a specific business challenge: Amazon Connect Decisions, Amazon Connect Talent, Amazon Connect Customer, and Amazon Connect Health. All four are designed to integrate into how your teams already work, not the other way around.

You may already know Amazon Connect. It began as the technology powering Amazon's retail customer service, and we've spent years learning how to run it at scale. That success taught us how to build AI solutions that work in the real world, and now we're applying those same learnings across other critical business functions where we have similar operational expertise. As part of this evolution, our original customer engagement solution, Amazon Connect, is now called Amazon Connect Customer.

These new Connect solutions draw on our expertise incorporating agents throughout Amazon's operations. We oversee a supply chain with more than 400 million SKUs. We hired 250,000 seasonal employees for the 2025 peak season alone. We handle millions of customer interactions every day. We support clinicians and patients through One Medical and Amazon Pharmacy. We've built AI systems not just to think about these challenges but to solve them in the real world, at scale, every single day.

Amazon Connect Decisions optimizes supply chains while humans stay in control

Supply chain disruptions take most companies more than two weeks to resolve, with many facing delays that cost millions in working capital, stockouts, and penalties. Teams spend days gathering data manually, coordinating across fragmented systems, and trying to translate demand forecasts into executable supply plans. When a supplier falls behind or a promotion overperforms, planners often spend days in spreadsheets and emails trying to understand what changed and who will be impacted before they can even decide what to do.

New product launch planning rule using P30 quantile to minimize excess inventory by $85K in first quarter

Amazon Connect Decisions is built on more than 25 specialized supply chain tools and 30 years of Amazon operational science, including one of Amazon’s Supply Chain Optimization Technologies (SCOT) foundation models. It starts with the goals your company wants to optimize and adapts to how your business runs.

Unlike traditional systems that require significant time and resources to configure, Connect Decisions provides both the forecasting models and the AI agents that do the heavy lifting. AI teammates understand your business context, set up the right forecasts for each individual product, and keep everything updated as your business evolves. They proactively ask for information that will impact forecasts—like upcoming promotions or holidays—and factor those inputs into results. At all times, they provide complete visibility and transparency into AI recommendations and decision-making, so you stay in control. AI teammates continuously learn from your team's actions, translating those learnings into better planning, analysis, and recommendations.

This intelligent approach is powered by sophisticated forecasting technology that adapts to any demand scenario. First, Connect Decisions applies the right mix of models to any demand pattern, from steady to bursty to brand-new products with no history. At its core, agents collaborate with localized ensemble forecast models, including foundation models like Chronos2, a time-series forecasting model, and one from SCOT, automatically selecting and tuning the optimal combination for each product.

When a manufacturing company launches a new automotive accessory with no sales history, Connect Decisions can link signals across similar product categories, analyze historical launch performance, and factor in regional demand patterns to generate a complete demand plan within minutes. When the planner adjusts a forecast or provides business context, the agent does what a good teammate would: captures that knowledge and applies it across similar products. What used to take weeks now takes minutes.

Second, Connect Decisions can replace thousands of alerts with a handful of actionable priorities that matter. When real-world disruptions hit—a supplier's fill rate drops, demand spikes in a region, or safety stock thresholds run low—Connect Decisions doesn't flood planners with noise. It does what a sharp analyst would do before walking into your office: triages the queue down to the top prioritized exceptions that need judgment, traces root causes automatically, and presents resolution options with projected impact, cost, and confidence scores. When a planner selects an option and explains why they chose it, the system treats that judgment the way a good teammate would, absorbing the reasoning and applying it going forward.

Every decision compounds: the AI gets smarter, the whole team gets better, and routine exceptions start resolving themselves. And because every plan gets smarter from execution feedback, and every decision gets better from planning context, Connect Decisions creates a compounding advantage that appreciates over time. Companies like Wells Vehicle Electronics and TVS Motors are already using Connect Decisions to transform their supply chain operations. Learn more about Connect Decisions.

Amazon Connect Talent accelerates hiring from weeks to days

Hiring at scale is something we've learned a lot about at Amazon. For example, during last year’s peak season, we hired 250,000 seasonal employees. Recruiters at any company can often start an average week with more applications than they’re able to work through, even if many of the candidates could be a great fit.

We see many organizations face the same challenges when it comes to high-volume hiring. Every unfilled role costs businesses momentum, revenue, and customer experience. For organizations hiring at scale—whether filling hundreds of seasonal roles, staffing new locations, or managing continuous recruitment—the challenge is even more pronounced. Recruiters juggle fragmented tools to post jobs, screen applications, and schedule interviews, all while racing to fill roles before hire-by dates. Poor job fit leads to early attrition, and delays send talent with the right skills to competitors who move faster.

Amazon Connect Talent accelerates the entire hiring process from job description to hiring decision. Starting with an existing job description, AI agents analyze the role requirements and generate a complete interview plan—identifying key competencies, creating structured questions, and building evaluation criteria. Recruiters review and approve the plan before any candidates are contacted. Once approved, the system automatically invites candidates to interview on their own schedule, 24/7. For candidates, this means a recruitment process that fits their life. No more trying to schedule phone screens during work hours or missing opportunities because of scheduling conflicts. Interview at 9 p.m. after putting the kids to bed, during a lunch break, or on a weekend—whenever works best. AI agents conduct voice interviews around the clock, asking consistent, job-related questions and adapting naturally to candidate responses.

AI interviewer asking candidate to describe a time they went above and beyond for a customer during a 4:28 minute interview call

Candidates can represent themselves through natural conversation rather than submitting an application into the void. The agent does the legwork overnight, so the recruiter walks in the next morning to a prepared brief with anonymized competency scores, full transcripts, and notes—not a blank page—before evaluating if the candidate is a good fit for the role.

Connect Talent is engineered to promote objectivity in the recruitment process by focusing on skills and competencies, not demographics. Candidate names and identifying information are stripped from recruiter dashboards, meaning recruiters see competency scores and performance data, not resumes. This skills-based approach matches the right candidate with the right skills to the right job, helping recruiters compare candidates objectively across hundreds of applicants and make data-driven hiring decisions.

AI agents evaluate candidates with job-related assessments and surface qualified candidates faster than traditional processes. Recruiters review AI-generated scores with full transcripts and notes. The agent surfaces its reasoning, and the recruiter makes the call. This enables recruiters to focus on relationships, not administrative work. The result is faster hiring, better matches, and a candidate experience that respects people's time. Learn more about Connect Talent.

Amazon Connect Customer brings proven capabilities to customer engagement

For years, companies like State Farm, Air Canada, and U.S. Bank have relied on Amazon Connect Customer, previously known as Amazon Connect, to deliver intelligent, personalized customer experiences across voice, chat, and digital channels.

Dashboard showing 60% churn reduction with 2-hour callback automation, displaying churn rate graph, callback times of 1hr 9min to 32min, and CSAT score of 4.2

When we expanded our fully agentic capabilities last year, customers were eager to go further: AI agents that understand when a customer calls about a delayed flight and proactively offer rebooking options, or real-time assistance that helps representatives resolve billing questions faster by surfacing relevant account history and next steps. Yet many organizations still struggle with legacy platforms that have been black boxes requiring proprietary professional services to integrate with today's AI systems, preventing businesses from experimenting quickly, improving their customer experience, or moving at the speed their customers demand.

Amazon Connect Customer now offers new configuration capabilities that enable organizations to set up conversational AI in weeks, not months, and configure experiences without technical expertise. Business teams can design and deploy sophisticated customer experiences directly, handling complex workflows from identity verification and payment processing to personalized product or service recommendations and issue resolution. Once deployed, they can keep iterating, make changes, and measure results immediately.

Enterprise customers including United Airlines have deployed sophisticated conversational AI experiences using these capabilities in as little as three months, going from initial concept to live production handling real customer interactions at scale, a process that traditionally takes six months or more. This rapid deployment without sacrificing quality or requiring extensive technical resources accelerates time to value from proof of concept to production. Learn more about Amazon Connect Customer.

AI that transforms how work gets done

Amazon Connect represents a new approach to building AI solutions for business. Rather than adding AI features to existing software, we designed these products from the ground up around a simple principle: AI should work like a teammate, not a tool.

We call this design philosophy humorphism. Traditional software interfaces were built around the metaphor of a desktop with folders, files, and buttons, which made sense when computers were tools you operated. But AI agents can reason, remember, and act independently. They need a different interface, one that mimics how people actually collaborate. When a colleague notices you're stuck on a problem, they speak up. When they see you're focused, they wait. When they learn something new, they apply it going forward. Amazon Connect products work the same way by learning your business context, adapting to how your teams work, and getting smarter over time.

Our set of Amazon Connect services delivers purpose-built agentic AI capabilities where they can have the biggest impact. They integrate into existing processes as AI teammates that elevate what your people can do and improve the outcomes that matter most to your business. And the agents get better and more useful over time as they have more context about your specific business, data, and workflows.

The capabilities we're introducing today join two other services we recently released focused on health: Amazon Connect Health for health care delivery and Amazon Bio Discovery for life sciences researchers, which tackle some of the biggest challenges facing health care companies and researchers.

We're just getting started with what Amazon Connect can do to help transform your business.