Every ticket across a family of lesson brands gets a note and a draft before a person opens it

Lesson Pro connects students with instructors for private lessons across the United States. Its best known brand is GOSKATE, which has taught skateboarding to beginners since 2009: over 120,000 lessons, more than 8,000 students, in three countries, with the instructor coming to the student's driveway or local skate park. The wider group runs lessons in bartending, archery, tennis, swimming, ice skating and more, and its parent HelloLesson lists more than 40 subjects.

Visit Lesson Pro at goskate.com →

One support team answers for many brands. A ticket about a wedding bartender and a ticket about a bartending course look alike in the subject line and belong to different businesses with different answers and different signatures. Getting that wrong in a reply is worse than replying slowly.

The booking side has its own repeated work. For every lesson sold, someone has to find an instructor near the customer, and the best source is the instructors who already taught nearby, which lives in old tickets. And each week the team needs to know which lessons still have no instructor, ordered by how soon they start.

Lesson Pro connected its Zendesk to eesel in June 2026 and added a second agent for the GOSKATE Zendesk in July. Both read the help center, past tickets and goskate.com, and both follow the same first rule: identify the brand before writing a word. Slack is connected too, for reports.

• A customer message wakes eesel on the ticket. It reads the thread, searches past tickets and the help center, leaves an internal note and drafts the reply. The team sends it. A small number of replies go out directly.
• Some tickets are skipped on purpose: job applications, voicemails, order forms and lesson sign-off forms are for people or other systems, not the agent.
• Brand identification comes from the ticket's company field, then from the lesson type. Replies carry the right brand's signature, and a GOSKATE reply is signed for the customer's city.
• Instructor search runs from an @eesel in an internal note. eesel maps the towns near the lesson, searches each one in Zendesk for completed and in-progress lessons, and lists the instructors it finds, with the tickets they taught on.
• Monday plans. Early each Monday, New York time, a scheduled run for each booking agent lists that agent's lessons still waiting for an instructor, soonest first and overdue ones flagged, and posts a plan for the week into a Slack channel.
• Weekly syncs pull new tickets and help center articles from both Zendesk accounts, so the instructor history it searches stays current.

Between 8 August and 8 September 2026 eesel worked about 1,900 Zendesk tickets for Lesson Pro, about 1,300 on the Lesson Pro account and about 580 on GOSKATE. It left about 1,700 internal notes, ran about 660 knowledge searches, drafted about 560 replies for the team and sent 26 replies itself. Most of the work lands between 10am and 7pm US Eastern time, and the volume has held between about 370 and 490 tickets a week since late July.

A third agent, Sales Bot, was set up in September for the sales team in Slack. A sales agent on a call @mentions it with a customer's town or ZIP code, and it checks both Zendesk accounts for instructors within about 30 miles and answers with first name, last initial, area, and what they have taught before, never contact details. It has not started work yet.