AI Inbound Voice Agent for Water Purifiers – Progetto Acque
Progetto Acque automates first contact with an inbound AI voice agent for water purifiers: 2,260 calls/month and 76.5% success. Want to free your team from repetitive requests?
Managing support for water purification systems means answering repetitive, yet urgent, calls. In this case study, we demonstrate how an inbound AI voice agent for water purifiers — DeepAgent's LARA — handled Progetto Acque's first contact, collecting complete data and providing immediate responses. In 30 days, it managed **2,260** calls with an operational success rate of **76.5%**, reducing wait times and freeing up the team for high-value interventions. ## TL;DR - **2,260** calls/month handled by the inbound AI voice agent with immediate 24/7 response. - **76.5%** operational success rate, **+31 p.p.** compared to the first month (45.5%). - **1,730** requests closed with the objective achieved and complete data for the team. - Intense adoption: **+10,173%** volume vs. initial phase (from 22 to 2,260 calls). - Filtering repetitive requests concerning water purifiers and providing ready information for ticketing. ## Context and Objectives ### Progetto Acque's Profile Progetto Acque operates in water treatment and purification, providing continuous assistance for systems and devices: filter changes, CO₂ refills, water softeners, water coolers, and administrative requests. Inbound calls are numerous, repetitive, and cover a wide range of technical and administrative topics. ### The First Contact Challenge The first level is critical: quickly understanding the need, collecting actionable information, and correctly routing requests. Without an automatic filter, the burden falls entirely on operators, leading to potential wait times, incomplete data, and less time for value-added activities. ## Solution: LARA, DeepAgent's AI Voice Agent ### What it does - Greets every call with an immediate response and continuous availability. - Recognizes intent: logs a case or answers informational questions. - Gathers necessary data to open or correctly route the request. - Promises and schedules callback when human intervention is needed. ### Why it works in the water sector An inbound virtual agent for call management related to systems and maintenance doesn't just "route"; it prepares clear, complete, and immediately actionable cases for topics such as filter changes, CO₂ refills, checks on water softeners and water coolers, malfunction reports, and administrative requests. ## Inbound Operational Flow When a customer calls, LARA receives the request, understands the context, and proceeds: - Records data and opens the case for technical interventions. - Directly answers informational questions (manuals, standard times, costs, procedures). - Routes to the correct channel or schedules a callback. Topics handled: - Filter changes and CO₂ refills - Maintenance and checks on water softeners and water coolers - Reports of system malfunctions - Administrative requests related to the service ## Implementation and Integrations - Telephony integration (SIP/VoIP) for immediate response. - Connection to CRM/ticketing to create and update cases. - Knowledge base for FAQs and operational policies. - Webhooks for notifications and callback workflows. Rapid rollout: go-live on **February 13, 2026**, with continuous improvement on intents and conversational datasets. ## Results and KPIs The agent has shown a clear consolidation trajectory: - **2,260** calls managed in the last 30 days (from **22** in the initial phase): **+10,173%** in volume. - Operational success rate at **76.5%** (vs **45.5%** in the first month, **+31 p.p.**). - **1,730** calls closed with the operational objective achieved. - **4,177** minutes of conversation managed. ### KPI Table (last 30 days) | KPI | Value | |---|---| | Calls Handled | **2,260** | | Calls with Objective Achieved | **1,730** | | Operational Success Rate | **76.5%** | | Conversation Minutes | **4,177** | | Volume Growth vs. Initial Phase | **+10,173%** | ### Interpreting the Success Rate Data The automatic technical flag indicated a **48.0%** success. A specific re-evaluation showed that many calls classified as "failure" were actually useful: data collected, case opened, callback promised. Considering the real objective (to manage first contact and prepare actionable requests), the success rate rises to **76.5%**.  ## Impact on Team and Customers - Less time on the phone for repetitive tasks; more focus on technical interventions. - More complete and standardized information for technicians. - Consistent service quality even during peak times, with reduced wait times. - End-to-end traceability of calls and cases. ## Why it's not a simple IVR - Natural language and intent comprehension, not just DTMF tones. - Structured and verified data collection during conversation. - Ability to answer FAQs and operational policies. - Smooth hand-off: promises and schedules callback when necessary.  ## AI Doesn't Replace Assistance, It Enhances It DeepAgent's LARA has transformed Progetto Acque's first-level support into continuous and precise management, relieving the team from peak loads and improving service KPIs. Do you want to see the impact on your inbound calls? Request a live demo: leave your contact information and try the AI agent. You will receive a callback from +39 06 8345191. ## Frequently Asked Questions ### How is the operational success rate of an AI voicebot for customer service measured? Clear objectives are defined for the first level (e.g., complete data collection, case opening, FAQ answers) and compared with conversational outcomes. Typical KPIs: operational success rate, calls closed on first contact, minutes handled, useful hand-off rate, and quality of data collected. Dashboards and sample audits ensure reliable interpretation. ### What integrations are needed for an inbound virtual agent in water utilities? At a minimum: SIP/VoIP telephony integration, CRM or ticketing for creating/updating cases, a knowledge base for FAQs, and webhooks for notifications. Optional: ERP for plant master data, technical calendars for callback, and payment systems where applicable. This way, call center automation for water utilities integrates into existing processes without disruption. ### How does an AI voice agent handle technical requests for water purifiers? It recognizes the intent (filter change, CO₂ refill, malfunction), guides the collection of critical data (model, system code, symptom), and verifies and confirms during the conversation. For informational questions, it draws from the knowledge base; for technical tickets, it opens a case and schedules a callback. The hand-off to operators or technicians occurs with already structured information. ### What are the implementation times and operating costs of an AI voice agent? A standard project requires 2–4 weeks: phone setup, intents, integrations, testing, and tuning. Operating costs are typically consumption-based (minutes/calls) with scalable packages. ROI comes from reduced wait times, increased first-contact resolution, and freed-up technical time. Scalability allows managing peaks without hiring additional staff. ### How is GDPR compliance and data security guaranteed? Data minimization, encryption in transit and at rest, access control, and configurable retention are applied. Call recording occurs with informed consent. EU data residency and Data Processing Agreements (DPAs) are available. Audits and logs allow full traceability. The inbound AI voice agent operates in compliance with industry policies and regulations. ### How is it different from a traditional IVR? An IVR requires menus and buttons; an AI voicebot understands natural language, asks targeted questions, and validates data. Besides routing, it opens cases, answers FAQs, and promises callback. The experience is faster and more precise, with measurable KPIs such as operational success rate and data quality, particularly useful for water purifiers.