Operanto — Operational Cockpit for Customer Operations

Context

Operanto is an internal Inovativi product, positioned as the operational cockpit for customer operations. It targets conversation-driven and customer-intensive businesses that run sales and service out of scattered messages, inboxes, and spreadsheets but lack a single place that holds operational context. Rather than replacing a CRM or acting as an autonomous agent, Operanto adds an operational memory and work layer around the systems a business already uses: it remembers customers and conversations, tracks commitments and operational context, connects work across channels and systems, and keeps a human in control of consequential decisions.

Problem

Conversation-driven teams run sales and operations on scattered DMs, inboxes, and spreadsheets. A message arrives, and turning it into a tracked commercial outcome — qualified, quoted, approved, scheduled, and pushed to a CRM — depends on manual coordination at every step. AI only helps if it acts inside a real, multi-tenant workflow with durable state, deterministic transitions, role-based control, human approval on sensitive actions, idempotent external calls, and audit trails.

What was built / modernized

Operanto turns each qualifying conversation into an Opportunity — the commercial object that gathers its conversations, extracted requirements, a workflow instance, quotes, appointments, approvals, and documents. The MediaSync layer ingests messages through real channel connectors (signature-verified webhooks, identity resolution, consent, and delivery status), an AI assistant scores intent, sentiment, and lead value and extracts structured requirements, and a config-driven workflow engine advances each opportunity through deterministic steps: collect requirements, draft a quote from the catalogue and business rules, gate sensitive actions behind human approval, book an appointment, and push idempotent, retried actions to CRM/ERP. AI proposes; deterministic software and roles control; people approve; every step is logged. Multi-tenancy is enforced at the data layer and RBAC is server-enforced.

Workflow highlights

• Real channel connectors (WhatsApp, Instagram, Messenger, Telegram, Viber, SMS) behind one contract, with signature-verified webhooks that degrade safely until credentials are set
• Opportunity spine: each qualifying conversation becomes a commercial object gathering requirements, quotes, appointments, approvals, and documents
• Config-driven workflow engine — verticals are data, not forked code — advancing opportunities through deterministic, audited step transitions
• AI tasks extract structured requirements, detect what is missing, draft quotes from the catalogue, and pull fields from photos and PDFs
• Human approval gates for sensitive actions (sending a quote, price overrides) and idempotent, retried CRM/ERP integration actions

Security, auditability & governance

• Multi-tenancy enforced at the data layer — every query scoped to the caller's workspace
• Server-enforced RBAC from a single permission matrix, extended to opportunities, quotes, approvals, and integrations
• AI proposes and humans approve — no customer message or sensitive action sends itself by default
• Every AI action logged with its prompt and confidence; workflow, quote, and approval transitions written to audit and transition records
• Idempotent, retried integration actions for CRM/ERP, mirroring the communication layer's webhook and sync pattern

Value delivered

• Customer messages become tracked opportunities — qualified, quoted, approved, and scheduled — instead of scattered DMs
• One config-driven engine runs multiple verticals (windows, solar, HVAC) as configuration rather than separate codebases
• AI accelerates qualification, quoting, and follow-up while humans, roles, and approval gates keep control
• Managers get one command center across conversations, opportunities, response times, and operator load

Technologies

Relevant roles