Case Studies

>

Spectora

Spectora Built 8 AI Agents from Real Call Transcripts to Ramp New Hires 2x Faster

Spectora is a home inspection software platform. It gives inspectors and inspection companies tools for scheduling, report writing, client communication, and business management.

Located

United States

Industry

SaaS

8

AI agents built from real call transcripts

2x

Faster new hire time to first live call

2

Separate agent libraries by team function

See How Outdoo Turns Practice Into Performance with AI Roleplays?

Book a Demo

Spectora's sales org runs two distinct functions: a new sales team acquiring net-new customers, and a growth team expanding revenue with existing accounts.

Kaitlyn Nichols manages sales enablement and the growth team. Stephen leads new sales. Together, they built a practice infrastructure that serves both functions without conflating them.

The Challenge

New sales reps joining Spectora needed to learn both the product and the conversation structure for selling to home inspectors, a buyer with specific concerns about ease of use, report quality, and the cost of switching platforms. Without structured practice, they were developing those skills on live calls where mistakes had real consequences.

Spectora had call recordings through Fathom, but those recordings were used for review, not practice. The best reps were already having excellent conversations with home inspector buyers. There was no systematic way to turn what those reps were doing into something new hires could practice against before taking their first calls.

The two-team structure created a second problem. New sales conversations, which involve outreach to inspectors unfamiliar with Spectora or actively considering a competitor switch, require completely different skills than growth conversations, which involve listening for expansion signals and navigating cost concerns with existing customers. A single generic roleplay scenario couldn't serve both.

The Solution

New sales agents built from actual Fathom transcripts

New sales reps at Spectora complete a series of roleplay sessions with the new sales agents as part of onboarding. The agents are configured to raise the objections, hesitations, and questions that real home inspector buyers raise. Common scenarios include inspectors locked into a competitor, those worried about the time required to switch platforms, and buyers asking specific questions about report quality and mobile functionality.

Scorecard feedback after each session gives new hires behavior-level data on what they did well and where they fell short. Managers use that data in coaching conversations during onboarding, addressing specific gaps before a rep takes their first live call.

Kaitlyn Nichols

“Building the agents from actual Fathom transcripts changed the quality of the practice completely. New hires are not practicing against a made-up buyer. They are practicing against a version of the real conversations our best reps are already having.

Kaitlyn NicholsSales Enablement and Growth Lead, Spectora

Growth reps get agents built for expansion, not new sales

The growth team uses a separate set of agents built specifically for expansion conversations. Those conversations have a different rhythm: listening for signals that a customer is ready to expand, navigating concerns about adding cost, and positioning additional features against value the customer has already seen.

Keeping the two libraries separate means growth reps practice skills relevant to their actual role rather than recycling new sales scenarios that don't reflect the expansion dynamic. As the team's call library in Fathom grows, the agents can be updated to reflect how the best expansion conversations are evolving.

Stephen

Expansion conversations are not just easier new sales calls. They require different instincts.Having agents that reflect that difference means the growth team is actually practicing what they do, not a close approximation of it.”

Stephen TucciRevenue Growth Lead, Spectora

Agent library stays current as best calls evolve

Because the agents were built from Fathom transcripts rather than generic templates, the practice environment reflects how Spectora's buyers actually behave. When scorecard data shows a new hire struggling with a specific objection type, coaching conversations can reference real examples of how that objection comes up rather than hypothetical ones.

The transcript-based foundation also means the library doesn't drift from reality. As the team's best calls evolve, the agents can be updated to match.

Kaitlyn Nichols

“Using our own transcript library as the foundation for the agents means the practice scenarios evolve as our best calls evolve. We are not locked into a static scenario that was accurate when we built it but drifts from reality over time.

Kaitlyn NicholsSales Enablement and Growth Lead, Spectora

The Impact

Since building the eight-agent library, Spectora has cut the time it takes new sales hires to reach a consistent performance baseline. Reps now enter their first live calls with multiple documented practice sessions behind them and scorecard data showing exactly where they need to focus.

New hires reach performance baseline faster

Representatives now complete multiple roleplay sessions with specific scorecard feedback before their first live call. The gap between a new hire's first day and a reliable performance baseline has narrowed because practice no longer depends on accumulating live call experience.

Eight agents covering two distinct conversation types

New sales agents and growth agents reflect different buyer dynamics and different required skills. Both teams practice scenarios built for their actual role, not scenarios borrowed from a function that works differently.

Practice grounded in real buyer behavior, not assumptions

Building agents from actual Fathom recordings means new hires encounter the same objections and conversation dynamics they'll face in live calls. The library stays current as the team's best conversations evolve, so the practice environment doesn't become stale.