For decades, the primary way travel agents developed selling skills was through mentorship. A senior agent sat beside a newer agent, listened to customer calls, provided feedback, and shared the techniques that worked for them. It was personal, relevant, and effective — when it happened.
The problem is that it rarely happens enough. Manager-to-agent coaching in the average travel agency occurs once a month at best. Many agents receive no structured coaching at all. The result: skill development is slow, inconsistent, and dependent on having the right manager with the right availability.
AI coaching doesn't replace the value of human mentorship. It replaces the scarcity of it.
The Mentorship Gap
Why Traditional Coaching Fails at Scale
The maths is simple. A team manager with 12 agents has approximately 40 hours per week. Customer management, operational duties, reporting, and administrative tasks consume 30+ hours. That leaves perhaps 10 hours for coaching — less than an hour per agent per week, assuming nothing interrupts.
In reality, CIPD research on UK sales teams shows:
- 35% of agents receive structured coaching less than once per month
- 22% receive no formal coaching at all
- Average coaching sessions last 12 minutes (versus the recommended 30-45 minutes)
- Coaching quality varies enormously between managers — some are excellent coaches, others are excellent sellers who struggle to articulate their own techniques
The agents who get coached improve. The agents who don't, plateau. The difference in performance between coached and uncoached agents is 20-30% on conversion rates — a gap that's entirely attributable to coaching, not talent.
The Distributed Workforce Challenge
The mentorship model breaks down further for distributed teams. Homeworking agents, franchise partners, and remote teams can't benefit from walk-the-floor coaching. A manager can't listen in on customer calls when the agent works from their kitchen table 200 miles away.
ABTA data shows that homeworking and remote selling now accounts for a significant proportion of UK travel agent distribution. These agents need coaching more than office-based agents — they have fewer opportunities for informal learning and peer support — but they receive less of it.
What AI Coaching Actually Does
AI coaching analyses agent selling performance and provides specific, actionable feedback. Here's what that looks like in practice:
After a Roleplay Session
An agent completes a roleplay scenario — handling a customer who wants a luxury honeymoon but has expressed concern about the price. After the 5-minute scenario, the AI analyses the conversation across multiple dimensions:
Product knowledge: "You recommended the Maldives Sunset Resort correctly for a honeymoon couple, but you described the water villa as 'having a glass floor' — it actually has a glass-bottom bath. This distinction matters when the customer has researched the property."
Selling technique: "When the customer said 'that's more than we wanted to spend,' you immediately offered a discount. A stronger approach would be to explore what 'more than we wanted to spend' means first. Try: 'I understand — tell me what budget you had in mind and I'll show you how we can create something special within it.' This preserves margin and shows flexibility."
Upselling: "You didn't mention the honeymoon package add-on (champagne, spa credit, sunset cruise) which has an 85% acceptance rate with honeymoon couples. At the Deciding stage, try: 'Most of our honeymoon couples add the romance package — it includes a private sunset cruise and spa treatment. Shall I include that in the quote?'"
Overall assessment: "Your customer rapport was strong — you asked about their wedding date and congratulated them, which created emotional connection. Focus on deepening your price objection handling and incorporating ancillary recommendations naturally."
This feedback is specific to the agent's actual performance, actionable, and available within seconds of completing the practice. No manager needed. Available at 6am or 11pm. Every time.
After an Assessment
When an agent completes a product knowledge assessment, AI coaching connects the results to selling implications:
"Your Caribbean assessment score was 62% — below the team average of 74%. The specific gaps are in Eastern Caribbean islands (Barbados, St Lucia, Grenada). Your last 3 Caribbean enquiries resulted in bookings for Jamaica and the Dominican Republic, which you know well, but you passed an Antigua enquiry to a colleague. Building Eastern Caribbean knowledge would let you handle those enquiries confidently. Suggested: Complete the Eastern Caribbean module and practise one roleplay scenario before your next shift."
Over Time
Unlike individual coaching sessions, AI coaching tracks patterns across weeks and months:
"Over the last 30 days, your objection handling scores have improved from 55% to 72% — particularly on price objections, where your responses are now much stronger. Your remaining development area is handling the 'I'll think about it' objection — you tend to agree and end the conversation rather than exploring the underlying concern. Focus your next 3 roleplay sessions on this scenario."
This longitudinal view is something most human managers can't provide — they remember the last coaching conversation but rarely have data on the progression over time.
AI Coaching vs Human Coaching
| Dimension | AI Coaching | Human Coaching |
|---|---|---|
| Availability | 24/7, after every interaction | Limited by manager time and schedule |
| Frequency | Multiple times per day if needed | Weekly to monthly at best |
| Consistency | Same standard for every agent | Varies by manager skill and mood |
| Specificity | Analyses exact words and techniques used | Relies on manager observation and memory |
| Scalability | Works for 5 agents or 5,000 | Limited to manager span of control (8-15) |
| Data integration | Connects coaching to performance data | Usually disconnected from performance data |
| Empathy | Cannot read emotional state or personal context | Understands personal circumstances and motivation |
| Relationship | Transactional | Builds trust and loyalty |
| Judgment | Pattern-based | Contextual and nuanced |
| Motivation | Provides feedback; can't inspire | Can motivate, encourage, and inspire |
The comparison reveals a clear pattern: AI coaching excels at frequency, consistency, specificity, and scalability. Human coaching excels at empathy, relationship, judgement, and motivation. The combination is more powerful than either alone.
The Combined Model
The most effective coaching model uses AI coaching as the foundation and human coaching as the enhancement:
AI Coaching (Daily)
- Provides immediate feedback after every roleplay session
- Tracks performance trends and identifies development areas
- Suggests specific practice scenarios based on identified gaps
- Maintains consistent coaching quality across all agents
Human Coaching (Weekly)
- Reviews AI coaching data to identify themes and priorities
- Addresses emotional and motivational factors that AI can't
- Provides contextual guidance (personal circumstances, team dynamics)
- Builds the manager-agent relationship that drives retention
The Manager's Role Changes
In this model, managers don't need to observe every call or assess every interaction — the AI handles that. Instead, managers focus on:
- Interpreting AI data: "The AI shows your objection handling improving but your needs analysis declining. What's changed in how you start conversations?"
- Addressing human factors: "I notice you've been less engaged with practice this week. Everything okay?"
- Strategic development: "Your Caribbean knowledge is strong, but we're seeing more Asia enquiries. Let's plan your development pathway for the next quarter."
- Recognition and motivation: "Your conversion rate hit 38% this month — highest on the team. What are you doing differently?"
Harvard Business Review research on sales coaching effectiveness shows that the highest-performing sales teams are those where coaching is frequent, specific, and data-informed — exactly the model that AI coaching enables.
Implementation
Step 1: Deploy AI Coaching Alongside Existing Practices
Don't replace human coaching — supplement it. Introduce AI coaching as a tool that helps agents develop between manager conversations.
Step 2: Train Managers on the Combined Model
Managers need to understand how to use AI coaching data. Show them how to read performance dashboards, identify coaching priorities from AI data, and structure their human coaching conversations around data-driven insights.
Step 3: Set Coaching Cadences
- AI coaching: After every roleplay session and assessment (agent-driven, daily)
- Manager coaching: Weekly 15-20 minute 1-to-1 (scheduled, non-negotiable)
- Team coaching: Monthly team session reviewing aggregate performance data
Step 4: Measure Coaching Impact
Track the relationship between coaching frequency and sales performance:
| Coaching Level | Expected Conversion Impact |
|---|---|
| No coaching | Baseline |
| AI coaching only (daily) | +15-20% |
| Human coaching only (weekly) | +10-15% |
| Combined AI + human coaching | +25-35% |
The combined model consistently outperforms either approach alone. McKinsey research on sales performance confirms that multi-channel coaching (combining automated and human elements) delivers 1.5-2x the impact of single-channel coaching.
The Mentorship Evolution
AI coaching isn't killing mentorship. It's making it possible at a scale that human resources alone could never achieve. Every agent in your business — whether they work in your flagship office or from their spare bedroom — can receive specific, frequent, data-driven coaching that helps them improve continuously.
Human mentors remain essential for the things AI can't do: building confidence, providing emotional support, sharing war stories, and inspiring agents to reach for their potential. The best mentors are those who use AI coaching data to make their human conversations more targeted, more specific, and more impactful.
The agents who develop fastest in 2026 aren't those with the best mentors or those with the best AI tools. They're those with both.
Explore AI coaching with TravAI →
This article is part of our AI in Travel & Tourism series. Related reading:
