How a business mentor took vacation without stopping student support
The situation
Joana leads an acceleration mentorship program for digital businesses, combining recorded training modules, validation assignments, and continuous individual support.
Alongside running the mentorship, Joana maintains a full-time corporate career as an employee. This dual responsibility imposes severe time constraints: all student support is concentrated at the fringes of the day, specifically early mornings, after 7 PM, and during weekends.
During regular business hours at her primary job, Joana is unable to access the training platform, answer student queries, or review assignments.
The operational bottleneck
With consecutive cohorts and a growing student community, support became the primary bottleneck in the business:
- End-of-day message backlog: At the end of each workday, Joana faced dozens of pending messages on the platform. She spent two to four hours each evening writing individual replies, resulting in physical exhaustion and sacrificing rest periods.
- Forced halts in student progression: When students encountered obstacles during the day, they remained stuck and unable to advance through exercises until Joana became available at night. This waiting period broke student execution momentum.
- High volume of repetitive queries: Data analysis revealed that approximately 70% of questions repeated concepts, technical steps, and guidelines already covered in recorded lessons or program documentation.
- Inability to take time off or disconnect: Taking a two-week vacation meant either suspending student support entirely or spending vacation time answering queries on a phone.
- Limitations of traditional hiring: Hiring a human support assistant entailed high fixed payroll costs and carried the risk of advice misaligned with the mentor's methodology.
The results
Designing and implementing an artificial intelligence automation infrastructure, powered by the mentorship's official documentation, restructured operations:
- 14 days of vacation without operational interruption: For the first time since launching the program, Joana was absent for two consecutive weeks without accessing the platform and without students experiencing any support gap.
- Sharp reduction in response time: Average response time dropped from over eight hours (the primary employment work interval) to under 45 seconds.
- Resolution of 78% of inquiries without human intervention: The assistant resolved almost eight out of ten questions on first contact, providing detailed instructions along with the lesson number and exact video minute where the concept is taught.
- Increased student capacity without added overhead: The subsequent cohort integrated 25 new students with no team expansion and zero additional working hours for Joana.
The process prior to implementation
The reliability and safety of a support assistant depend on the information engineering conducted before connecting technical tools. Pre-implementation work was structured in four phases:
1. Audit and cataloging of question history
An exhaustive extraction of over 300 messages and inquiries submitted by students from previous cohorts was conducted. This dataset was organized into four objective categories:
- Operational and platform queries: Assignment deadlines, material locations, and form submissions.
- Technical tool configuration: Resolving software and setup blocks in applications taught during the lessons.
- Methodology application: Clarifications on completing templates, spreadsheets, and scripts provided in the program.
- Custom strategic evaluation: Validation of commercial proposals or individual positioning decisions.
This classification established an objective operational boundary: automate the first three categories entirely, reserving manual intervention solely for personalized strategic guidance.
