Amazon Connect Expert Needed for AI-First Omnichannel Contact Center
We are looking for an experienced Amazon Connect implementation partner to deliver an AI-first, AWS-native omnichannel contact center in Amazon Connect (Region: eu-central-1 Frankfurt).
The work is primarily configuration and integration, plus delivery of production-ready flows, reporting, and handover documentation.Key constraints:
- AWSnative only (no vendorowned platforms or proprietary thirdparty contact center layers)
- Integrates with our bespoke systems (SSO/Identity + SOAP APIs)
- Payments tied to productiondelivered milestones
Chat must support live and asynchronous conversations, be AI-first with human escalation, and be multilingual based on the user's language setting.
Social includes WhatsApp and Instagram/Facebook (Meta Business). Email must support personal and team mailboxes with agent/supervisor visibility, deduplication, and ownership tracking. Outbound calling may be enabled, but outbound campaigns are explicitly out of scope.AI-first model: we require:
(1) a generic AI assistant trained on approved documentation (Confluence synchronized to S3 for Bedrock ingestion) and
(2) a customer-specific assistant with controlled access to account/campaign data based on user permissions. There must be a clear separation between general knowledge and customer-specific data. Human handoff triggers include explicit user request, repeated fallback, negative sentiment, or business-rule constraints. "AI-only" customers should be blocked from human escalation by default, with a billable exception flow.
Identity and data handling:
logged-in users are identified via userID from our platform; unauthenticated users via phone/email/profile details. All interactions must be linked to userID and customerID, and authorization must follow our permission model and tokens.
SOAP integration is required for read/write using ephemeral interaction-scoped storage with TTL, and only normalized result sets are allowed (no free-form queries).
Reporting and quality:
enable Amazon Contact Lens (sentiment + transcripts), use Connect built-in reporting/dashboards (volume, SLA, handoffs, CSAT, sentiment), and store transcripts/metadata securely with optional redaction.
Must-deliver outcomes:
- Amazon Connect instance fully configured (queues, routing profiles, hours, skills, languages, IVR with PIN verification). Instance ARN is available.
- AIfirst flows for onboarding and case handling: identify user, retrieve profile via SOAP (or Customer Profiles), convert natural language into SOAP write requests for allowed updates.
- Omnichannel setup: chat + WhatsApp + Instagram/Facebook with multilingual UX and context preserved across channels.
- Email ingestion: personal + team mailboxes, deduplication + ownership tracking, supervisor oversight, reply in same channel.
- Permissionbased routing and data access, with full context transfer on human handoff (profile, history, AI summary, entitlement, relevant platform attributes visible in the Agent UI).
- Contact Lens enabled, transcripts/metadata stored, operational dashboards configured.
- Documentation and handover: runbooks, scaling guide, security considerations, operating procedures.
Proposal format (please follow):
- Company & team: roles, relevant Amazon Connect references (IVR/PIN, omnichannel, Contact Lens, Bedrock/AI orchestration)
- Technical approach: AWS architecture, SAML SSO + SOAP integration security, permission enforcement, prompt injection/abuse controls
- Plan & responsibilities: milestones + timeline, and a clear split of what you deliver vs what you need from our stakeholders and developers
- Pricing & payments: fixed vs variable, optional items priced separately, milestonebased payments tied to production deliverables
Selection will prioritize proven Amazon Connect delivery (IVR/PIN, omnichannel, Contact Lens), secure enterprise integrations (SAML + SOAP preferred), AI safety and cost control, and strict AWS-native delivery.
Contract duration of 1 to 3 months.
Mandatory skills:
amazon connect, Chatbot Development, Amazon Bedrock, Amazon Contact Lens, AI Agent Development, AI Development, AI Bot, PostgreSQL, Amazon DynamoDB