DHL Freight
Autonomous genAI customer service chatbot and voicebot
VIVA, DHL Freight's AI assistant delivered as a chatbot and voicebot, answers 93% of customer requests without human intervention, deployed in Sweden (2021), France (2024), and the Czech Republic (2025).
Objective
Absorb recurring freight customer service requests (shipment tracking, claims, product questions) autonomously and 24/7, across several countries, without scaling up headcount.
The deployment
VIVA is the AI assistant for DHL Freight customer service, delivered as a chatbot and a voicebot under the internal ICE 2.0 project. A customer can track a shipment, open a claim, or ask a question about DHL Freight products, in writing through the help center or by phone with the voicebot, 24 hours a day. The 2025 version adds short-term memory that lets the assistant remember earlier questions within the same conversation. DHL first launched VIVA in Sweden in 2021 as an intent-based version, then moved to a genAI version able to phrase complete answers.
Results Proof C
Official communication from the DHL Freight brand (blog and VIVA product pages by country) with figures, and a multi-market deployment verifiable through the online help centers. Single organization source (DHL) but confirmed by several official pages and the existence of the services in production, hence C rather than B.
How it works
Documented architectureThe stack in detail
- infra ICE 2.0 (plateforme interne DHL Freight) conversational foundation of the internal program that carries the VIVA chatbot and voicebot in each country
- llm LLM genAI de VIVA generative version of the assistant since 2024, with short-term memory since 2025; the model vendor is not publicly disclosed
- outil Brique voicebot VIVA's phone channel, rolled out country by country alongside the help center chat
- infra API de suivi d'expedition DHL Freight real-time tracking data, the bulk of the request volume; if this link goes down, the assistant switches to a human
How it runs, concretely
For ops teams-
1Customer contact customer
The customer opens the help center chat or calls and states the request (tracking, claim, product question).
-
2Understanding AI
VIVA identifies the intent, with memory of earlier exchanges in the conversation since 2025.
-
3Autonomous answer AI
The assistant phrases a complete answer from tracking data and the product base.
-
4Escalate if needed AI
The 7% of out-of-scope cases are transferred to a customer service agent.
Access to shipment status and the product knowledge base. If the link to tracking data goes down, the assistant can no longer answer tracking requests, which make up most of the volume, and hands off to a human.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- real-time shipment tracking API
- product and procedure knowledge base
- conversation history for training
Org prerequisites
- a pilot customer service team per country
- multilingual governance and phased rollout
- a GDPR / AI Act framework for a customer assistant
Possible stack
- an LLM with intent orchestration
- a voicebot component for the phone channel
- connectors to the tracking system
The plan, step by step
- Step 1Frame the pilot scope (shipment tracking first) and the GDPR / AI Act setup (terms of use, privacy notice).Deliverable: Functional scope and legal framework published
- Step 2Connect the assistant to the tracking API and the product knowledge base.Deliverable: A bot able to answer tracking in test
- Step 3Launch the chatbot in a pilot country with escalation to an agent, and measure the autonomous resolution rate.Deliverable: Chatbot in production with a tracked resolution rate
- Step 4Add claims, product questions, and conversation memory.Deliverable: Broader functional coverage in the pilot country
- Step 5Add the voice channel and replicate the foundation in the next country.Deliverable: Voicebot live and a second country rolling out
First step: Deploy the chatbot on shipment tracking in a pilot country, measure the autonomous resolution rate, then add claims and the voicebot.
The same pattern, proven in other industries
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Bank of America
conversational financial assistant at scale
Sources
- S1 AI Chatbot VIVA Redefines Customer Service Interested party archive pending
- S2 AI-Based Customer Service Assistant Viva (help center Suede) Primary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.