Which KPIs should be measured in an outbound campaign?

Measure attempts, reachability, completed conversations, percentage of calls over 2 minutes, scheduled appointments, conversion per segment, and average time per site visit. Exclude inbound callbacks from outbound KPIs. These indicators show where to intervene on segmentation, scripts, and objection handling.

What changes compared to a traditional call center?

Every conversation is tracked and can be re-listened to, not just the outcome. Scripts and objection handling are optimized based on what truly converts. The goal isn't to sell over the phone, but to present the free solar site visit as a logical next step. Consultants receive more prepared appointments with context.

Should prices and technical details be discussed over the phone?

No. The phone call serves to understand the case, validate interest, and schedule a visit. Quotes and technicalities are evaluated during the site visit, reducing friction and misunderstandings. This approach increases data quality and consultant focus, maintaining a high appointment acceptance rate.

AI Voice Agent for Solar: 138 UltraSolar Site Visits

Q: How does the agent integrate with CRM and calendars?

The agent connects to the CRM to read and update statuses and notes and can propose availability windows, synchronizing the consultant's calendar. Unlike traditional solar appointment setting, everything remains tracked: outcomes, durations, and conversions by segment, allowing for continuous data-driven optimizations.

Q: Which lists work best for solar?

The best results come from the CRM segment, with known contacts showing signs of interest. However, volume comes from cold leads. The combination maximizes appointments and learning: qualified solar leads require targeted questions, while for cold leads, rhythm, timing, and objection handling are crucial.

An AI voice agent scheduled 138 solar site surveys in 60 days: segmentation, conversations over 2 minutes, and seamless handover to consultants. How to replicate this model with UltraSolar?

In the solar industry, contracts aren't closed over the phone. They're closed after a technician has inspected the roof, exposure, and electrical panel. That's why the crucial call is the one that gets the local consultant to the client's home. UltraSolar entrusted this call to Gaia, an AI voice agent for solar built on DeepAgent: in 60 days, 19,362 attempts, 5,249 completed conversations, and **138 scheduled site visits**. The breakthrough? Calls lasting over two minutes – few, but decisive. !AI Voice Agent for Solar: 138 UltraSolar Site Visits — figure 1 ## TL;DR - 19,362 attempts, 5,249 completed conversations, and **138 site visits** in 60 days. - Calls >2 minutes account for **3.73%** but generate **66.7%** of site visits. - Average conversion **2.63%**; CRM segment **7.73%**; cold leads = **73%** of volume. - One single KPI: scheduled site visit. Scripts and objection handling adapted for 3 segments. - End-to-end tracked outbound: an alternative to traditional solar appointment setting. ## The Client: UltraSolar ### Who They Are UltraSolar designs and installs turnkey photovoltaic systems, complete with storage batteries, monitoring, and after-sales assistance. A certified Solarwatt partner, they work with private individuals, condominiums, and businesses in central-northern Italy. Operational headquarters in Latina, logistical hubs in the Milan area, Bologna, and Rome. Their commercial model revolves around local consultants. ### The Starting Point Their CRM contained various lists: recently acquired leads, already qualified contacts, and people informed months prior. The obstacle wasn't the contacts, but time. Manually working through all of them would take hours away from consultants who need to be on rooftops, not on the phone. ## The Challenge: One Goal, Lists with Different Temperatures The objective was singular and non-negotiable: to schedule a free site visit with the local consultant. No phone sales, no pricing: just the necessary information for the visit and a clean operational handover. The difficulty was making this work across very different audiences while maintaining consistent messaging, data collection, objection handling, handover, and KPIs. ## How the AI Voice Agent for Solar Works, Phase by Phase ### 1) Scope and Segmentation Only one expected outcome: a scheduled site visit. Three outbound agent configurations for: already qualified leads, cold CRM leads, and a dedicated CRM segment. Same objective, but opening and arguments adapted to the context. This is where qualified solar leads come in: they require a different tone and questions compared to cold leads. ### 2) Campaign Activation List upload, launches per segment, reachability verification, and tracking of every attempt: call completed or not, duration, outcome. Everything remains measurable and comparable by segment. ### 3) Conversation and Qualification Gaia initiates the conversation, verifies interest, and maps the real situation: property type, existing system or not, energy needs. Then, it addresses typical concerns about technical feasibility, timelines, and installation reliability. ### 4) Closing and Operational Handover The site visit is proposed as a concrete step to evaluate the case. The agent collects availability and contact details, confirms the appointment with the local consultant, and records the outcome in a legible format. ### 5) Separate Callbacks Those requesting a callback are managed through a dedicated inbound channel, outside of outbound KPIs. This is a methodological choice to compare segments cleanly and transparently. > Key difference from solar appointment setting via call centers: every conversation is tracked, can be re-listened to, and analyzed; the best ones become material for improving scripts and objection handling. !AI Voice Agent for Solar: 138 UltraSolar Site Visits — figure 2 ## Results in 60 Days | Metric | Result | |---|---| | Call Attempts | **19,362** | | Completed Calls | **5,249** (27.1% of attempts) | | Free Site Visits Scheduled | **138** | | Conversion Rate on Completed Calls | **2.63%** | | Calls ≥ 2 minutes | **196** (3.73% of completed) | | Site Visits from Calls ≥ 2 minutes | **92** (66.67% of total) | | Conversion Rate of Calls ≥ 2 minutes | **46.94%** | | Conversion Rate of Calls < 2 minutes | **0.91%** | | Ratio ≥2 min vs <2 min | **51.6×** | | Average Site Visits per Week | **~16** | In essence: approximately one site visit for every **38** completed conversations, providing a constant pipeline for local consultants. Inbound callbacks are excluded from KPIs. !AI Voice Agent for Solar: 138 UltraSolar Site Visits — figure 3 ## The Key Metric: Real Conversations Out of 5,249 completed calls, only **196** lasted over two minutes. And from those, **92** site visits were generated: almost one in two ended with an appointment. The other **5,053** shorter calls generated **46** visits (less than 1%). A conversation exceeding two minutes converts over **50×** more effectively. This is where the free solar site visit stops feeling like an “offer” and becomes a logical solution to the client's problem. ## Segment Comparison: Volume and Quality - CRM Segment: The most efficient. Converts **7.73%** of completed calls (≈ **2.9×** the campaign average). - Cold Leads: The volume driver. **101/138** site visits (**73%**) originated here, with a **2.16%** conversion rate. Quality and quantity complement each other: the most efficient segment provides density, while cold lists ensure continuity. Without the latter, the pipeline would remain at a few dozen visits. ## The Conversation That Explains the Method The best call of the period started with a problem, not a new system. The client already has a solar installation that's underperforming, and the installer isn't responding. Gaia listens, asks questions, clarifies what's truly needed. Then, it proposes the only sensible thing: sending a technician to inspect. Appointment confirmed. Recurring themes in converting calls: evaluation of a new system or energy solution, issues with existing systems, reliability and assistance, technical and economic feasibility, availability for a visit. ## What to Learn (and Where to Improve) - One KPI: scheduled site visit. Everything else is a means, not an end. - Segment before calling: warm, cold, and CRM leads require different openings for the same goal. - Extend good conversations: getting more people past the **2-minute mark** is the most powerful lever. - Study the best calls: qualification questions and objection handling are refined from winning cases. - Keep channels separate: inbound and outbound measured separately make the numbers legible. > Want examples in other sectors? Read the Fincontinuo case study (outbound pre-qualification) and the credit case study (loan request callbacks). ## Conclusion The value of an AI voice agent for solar isn't “making more calls.” It's about transforming stagnant lists into real appointments for those who need to get on the roof. For UltraSolar: **138 site visits** in two months, two-thirds of which originated from conversations where the agent listened before proposing. Want to see how it works with your CRM? Request a demo. ## Frequently Asked Questions ### How does the agent integrate with CRM and calendars? The agent connects to the CRM to read/update statuses and notes and can propose availability windows, synchronizing the consultant's calendar. The advantage over traditional solar appointment setting is end-to-end traceability: every outcome is recorded, re-listenable, and measurable by segment, duration, and conversion. ### Which lists work best for solar? The best results come from the CRM segment (known contacts with signs of interest), but volume comes from cold leads. Combining both maximizes site visits and data. Scripts need to be adapted: qualified solar leads require different questions, while for cold leads, rhythm and objection handling are key. ### Which KPIs to measure in outbound? At a minimum: attempts, reachability, completed conversations, average duration and percentage of calls over 2 minutes, scheduled appointments, and conversion per segment. Filtering inbound callbacks from outbound KPIs avoids distortion. Comparing these indicators helps optimize targeting, openings, and objection handling. ### What's different compared to an external call center? All conversations are tracked and analyzable, not just the outcomes. The script is improved by listening to what truly works. The goal isn't to sell over the phone, but to propose the free solar site visit as a sensible next step. Consultants receive more prepared appointments with context. ### Should we discuss prices and technical details on the phone? No. The phone call serves to understand the situation, validate interest, and schedule the visit. Technical matters, quotes, and constraints are evaluated during the site visit. This approach reduces friction, increases data quality, and focuses consultants on high-value activities. Talk to sales → · Create your agent →