The Transportation Security Administration has deployed Ace, an AI agent that now handles 100,000 routine traveler conversations each month and resolves 96% of inquiries without human intervention. The system went live during the summer travel surge and represents a shift in how the agency manages passenger questions at scale - lowering costs, reducing contact center pressure, and freeing staff for security work.
Nearly 3 million people fly through U.S. airports daily. Questions about packing liquids, checkpoint procedures, or traveling with medical devices can slow passengers and pull officers away from screening duties. Ace answers those questions before travelers reach the airport, using a secure platform built on Salesforce Government Cloud with Missionforce.
What Ace handles day to day
The AI agent fields routine inquiries about checkpoint preparation, prohibited items, and special accommodations. When a question falls outside its scope, Ace routes it to human staff. The TSA's Customer Experience Branch led the deployment as part of a broader modernization effort to make traveler information more accessible and consistent.
"Nearly three million people travel through U.S. airports every day, and a simple question shouldn't stand between them and a smooth trip," said Kendall Collins, CEO of Missionforce and Government Cloud. "TSA's new Ace AI agent gives travelers fast, accurate answers before they reach the checkpoint - from what they can pack to how they can prepare - while helping TSA officers stay focused on their essential security mission."
The numbers behind the rollout
The TSA projects savings across three areas. First, the agency expects to automate 70,000 transactions annually that previously required 10 minutes of manual data-sharing each - reclaiming over 11,660 staff hours per year. Second, consumer interactions that once cost more than $4 per conversation have dropped by over 90% under Ace, contributing to an estimated $11 million in labor-based value over three years. Third, work that previously demanded the equivalent of 12 full-time employees to manage backlogs now runs through the AI agent, letting staff focus on high-priority cases.
"If we can help the traveling public easily access accurate, timely information, we can help alleviate stress in the field on travel days and increase the TSA's overall effectiveness," said Niki French, Customer Experience Branch Manager at the TSA.
Not a standalone chatbot
The TSA built two foundational applications before layering on Ace. The Customer Service Managers App, running on Service Cloud, unifies passenger inquiries from email, phone, and social media into a single view. Staff can review conversation histories, pull in subject-matter experts, and assign follow-ups without manual back-office requests. The TSA Cares App supports travelers with medical conditions or other circumstances requiring additional checkpoint assistance, connecting online requests directly to the contact center and regional coordination teams.
These applications create a centralized repository of traveler interactions. With Salesforce Shield providing security and monitoring, Ace draws on that unified data to deliver answers based on current TSA guidance. The phased approach means the AI agent isn't guessing - it's working from the same information human agents use.
For support teams evaluating similar deployments, the TSA's structure offers a practical reference point. The agency didn't start with AI. It started with data unification, then layered on AI Agents & Automation once the foundation was in place. That sequencing matters. An AI agent answering questions from fragmented or outdated sources creates more work, not less.
"The TSA is leading with trusted AI to meet the mission," said Paul Tatum, EVP of Global Public Sector Solutions at Salesforce. "They are delivering a secure, data-driven, and AI-first foundation that scales operations, lowers costs, and protects public trust, even during the most stressful travel times."
Why this matters for customer support professionals
The TSA's deployment challenges a common assumption that AI in support means replacing people. Here, the opposite happened. By automating 96% of routine inquiries, Ace didn't eliminate roles - it absorbed the overflow that previously required 12 full-time equivalent positions just to manage backlogs. The human team now handles complex cases, exceptions, and travelers who need individualized help. For support leaders, the takeaway is specific: measure what your team is spending time on that a system could resolve, calculate the hourly cost of that work, and compare it against the cost of inaction during peak periods. The TSA's $4-to-under-40-cents per conversation shift came from knowing exactly what routine meant in their environment before automating it.
The broader lesson from the TSA's approach to AI for Customer Support is that deployment order determines results. Unified data first, automation second, AI agent third. Teams that skip straight to the agent often spend more time correcting bad answers than the automation saves.
Your membership also unlocks:
