DVLA: Contact Centre Chatbot service
A Chatbot tool used by Contact Centre Webchat channel to automatically answer customers and collate relevant information for the Webchat advisor when required.
Tier 1 Information
1 - Name
DVLA Contact Centre (CC) Chatbot service
2 - Description
DVLA Contact Centre (CC) Chatbot service
Purpose: The Chatbot is used on the CC’s Webchat channel to automatically answer customers and collate relevant information for the Webchat advisor when required.
How it’s Used: Customers can get in touch via the Contact Us page on the DVLA web page on GOV.UK. The customer is then told they are through to DVLAs Virtual Assistant.
The chatbot uses AI (via Natural Language Processing) to answer simple frequently asked questions without the need for a Webchat advisor. The responses the chat bot provides are configured by DVLAs chat bot development team and customers can free key their enquiry into the bot and the NLP engine is used to understand the customer intent and route the customer to the pre-configured responses. The chat bot is integrated with the Contact Centre’s web chat solution, meaning customers can choose to speak to an advisor at any time. When this happens, the bot will ask the customer to provide relevant information to help the advisor answer their enquiry.
Data Input: The data is obtained directly from the customer as the customer will complete the pre-chat form and will provide information about their enquiry within the chat bot
Decision Support:
Why it’s Used: The use of Chatbot provides the CC agent with a set of answers designed to diagnose where a customer has encountered problems when using our online services. This means the agent can reduce their chat handling time as they don’t need to carry out the diagnostic questioning when the customer reaches them and can concentrate on assisting the customer with their transaction/enquiry quickly.
Resource Allocation: This particular chatbot has been in use since 2022 and is business as usual. DVLA has used chatbots since 2018
3 - Website URL
https://contact.dvla.gov.uk/driving-licence
4 - Contact email
Contact Centre Management Team ContactCentreCCMT@dvla.gov.uk
Tier 2 - Owner and Responsibility
1.1 - Organisation or department
Driver and Vehicle Licensing Agency
1.2 - Team
Contact Centre
1.3 - Senior responsible owner
Head of Contact Centre
1.4 - External supplier involvement
Yes
1.4.1 - External supplier
Content Guru (Supplier of the Storm Platform, who integrated the chat bot product with the web chat channel) Google DialogFlow (the chat bot product being used)
1.4.2 - Companies House Number
Content Guru - 05653869
1.4.3 - External supplier role
Content Guru are the supplier of our Cloud Contact Centre platform (Storm) which is integrated with the Google DialogFlow Chat bot product. When we migrated over to the Storm platform, Content Guru migrated our existing chat bot flows and scripts over to DialogFlow (we previously used Salesforce Einstein). Following this migration, all chat bot development was completed by in-house DVLA administrators
1.4.4 - Procurement procedure type
Open procurement using the Crown Commercial Framework - Network Services RM3808
1.4.5 - Data access terms
Content Guru provide Contact Centre Cloud services to the DVLA and are only provided permission based access to data as part of incident investigation and resolution. The chat bot solution processes customer enquiry information and captures some personal information to support the data protection confirmation process
Tier 2 - Description and Rationale
2.1 - Detailed description
Customers can get in touch via the Contact Us page on the DVLA web page on GOV.UK or the customer can drop out of a service and ask for assistance if they are experiencing issues.
The Chatbot effectively acts in the same way as an Interactive Voice Response (IVR) does on the telephony channel but on a website channel. It provides an upfront menu of enquiry options for the customer to select from and carries out a number of actions including: - Asking the customer questions to gather information. - Answering general enquiries with pre-scripted responses. - Routing customers to the appropriate Webchat advisor.
The data is obtained directly from the customer as the customer will complete the pre-chat form on the website when first engaging the chatbot. But instead of entering the queue to an advisor they will be routed to the Google AI Chatbot if the service is supported by one.
The chatbot uses AI to answer simple frequently asked questions without the need for a Webchat advisor.
There are 2 types of chatbot used, depending on the service. Natural Language Processing (NLP) Bot Interactive Voice Response (IVR) Bot
Services utilising the NLP bot are: 1. Vehicles General / Electronic Vehicle Licensing 1. Drivers General.
The Chatbot will be used to gather information from a customer via the webpage to identify the reason for their enquiry and the personal data required for DPA security checks. The Chatbot will then transfer the customer to a Webchat Advisor who will be able to see the chat transcript and the information gathered by the Chatbot.
IVR bots are used on the following services: 1. Driving Licence Online (DLO) 1. Drivers Medical 1. View Drivers Licence 1. Vehicle management & Personal Registrations 1. Unlicensed Vehicle 1. Trade Licensing Service
IVR bots are menu based. Menus are presented to the customer and they click the relevant response.
The Chatbot will be used to answer general enquiries from the customer without the need for advisor intervention.
2.2 - Scope
Chatbot / Webchat data is used to assist customers in using DVLA’s service.
The Chatbot is part of our Contact Centre platform and uses a feature called Google DialogFlow. The Chatbot uses NLP to understand customer intent and route to the relevant response and is used on the CC’s Webchat channel to automatically answer customers and collate relevant information for the Webchat advisor when required.
The use of Chatbot provides the CC agent with a set of answers designed to diagnose where a customer has encountered problems when using our online services. This means the agent can reduce their call time as they don’t need to carry out the verbal diagnostic questioning when the customer reaches them and can concentrate on assisting the customer with their transaction/enquiry quickly.
2.3 - Benefit
Benefits to the Agency, individuals and third parties include: 1. Automation of around 20% of customer enquiries on the web chat channel through bot generated auto responses. 1. Provide 300,000 customers per month with 24/7 online support. 1. Free up resource for more complex issues. 1. Improved customer service due to customer not needing to spend time queueing to chat to an advisor. 1. Reduction in an advisor’s chat handling time of around 2 minutes linked to chat bot data capture (minimising the advisor time asking for this).
2.4 - Previous process
Prior to the delivery of the online chat bot, customers were sent immediately to a web chat advisor queue meaning there was no automated option for customers to use to answer their queries without human support.
2.5 - Alternatives considered
No other options were considered, chat bots were the only solution which would provide automated responses based on customer intent and would integrate seamlessly with the advisor web chat channel.
Tier 2 - Decision making Process
3.1 - Process integration
DVLA Chatbot asks questions and provides answers based on programmed rules and logic. It decides which prepared response to provide based on how similar the customers inputted question / statement is to the expected input. If the chatbot cannot confidently interpret a question / statement respective of the question it was asked, it will transfer the customer a DVLA webchat advisor to continue the conversation.
The webchat advisor has access to DVLA databases so if any record access is required to answer the customer’s enquiry, then only the advisor can do this. The chat bot can only answer higher level general enquiries, record access and more complex enquiries require a human.
There are 2 main ways a customer would access our chat bot - From a DVLA digital service (like taxing a vehicle). The webchat/chatbot link is provided when something goes wrong to provide transactional support. -From the Contact DVLA page on GOV.UK as a general enquiry option when a customer can’t find what they need from the website. Customers are able to choose from telephone number, email service or web chat when they need to contact the DVLA.
3.2 - Provided information
The advisor sees the full conversation between the bot and the customer when a chat is connected (same interface). The advisor will also see an ‘interaction summary’ pane which outlines the information the bot has captured. This varies across services but will include info like: Enquiry reason Name Address Vehicle Registration Mark Make and Model of Vehicle Date of Birth
3.3 - Frequency and scale of usage
The Chatbot is used on a daily basis as part of business as usual process within the Contact Centre. Around 300k customers access the chat bot every month.
3.4 - Human decisions and review
NLP Bots The Chatbot will be used to gather information from a customer such as the reason for their enquiry and the personal data required for Data Protection security checks. The Chatbot will then transfer the customer to a Webchat Advisor who will be able to see the chat transcript and the information gathered by the Chatbot. The advisor will then verify the customer via the data protection security check information provided to complete the enquiry. If the customer does not pass the check, the information will be taken by the agent and checked again, if they do not pass for a second time, they will be advised to ring the Contact Centre to speak to a telephone agent.
Menu based Bots Menus are presented to the customer and they click the relevant response, these bots do not rely on NLP
The Chatbot will be used to answer general enquiries from the customer without the need for advisor intervention. If the customers query has been answered, they will exit the Chat but if record access is required to answer the customer’s enquiry, the Chatbot will transfer the customer to an advisor and the above data protection security check process will be followed and a human engaged.
3.5 - Required training
The chatbot has been designed to be used without the need for formal training. Customers are asked to provide answers to questions only.
All Contact Centre web chat advisors are trained on using the web chat channel and the features available to them. There are also training materials, knowledge articles, security documents and web chat response templates available to them.
The ‘start’ page for web chat and chat bot explains to the customer what they can expect from the channel, what the bot can and can’t answer, the opening hours of advisor web chat. It is made clear to the customer that initially they are speaking to a virtual assistant (bot) and can speak to an advisor at any time (escape hatch).
3.6 - Appeals and review
The Chatbot will not be making any decisions which will have a significant impact on the customer. The role of the Chatbot will be to gather information, offer general advice and to route to the appropriate advisor where applicable.
Ongoing maintenance and updates to the chat bot are carried out on a daily basis by the Contact Centre eChat bot administration team, they are continually checking for accurate routing and ensuring the NLP is sending the customers to the correct advice. The bot does not use any generative AI capability so only offers responses that this team have configured
The intended outcome of the processing will be that the customer gets their enquiry answered via the Chatbot, where this is not possible it will be routed to an advisor either via the Webchat or email channels.
Tier 2 - Tool Specification
4.1.1 - System architecture
The Chatbot is part of DVLAs Contact Centre platform, a feature called Google AI Machine Agent is used (Also referred to as Google AI DialogFlow).
The system architecture of Google AI Dialogflow is designed to support scalable, intelligent, and responsive conversational agents. It integrates multiple components such as natural language understanding (NLU), intent recognition, context management, and fulfilment services.
The high-level architecture: 1. User Interface (Input Channels) IVR Systems (for phone-based interactions) Dialogflow receives text or speech input through these channels.
-
Dialogflow Core Components NLU Engine - Parses user input to identify intents and entities using machine learning and NLP. Intent Classification - Determines the goal of the user’s query. Entity Extraction - Extracts relevant parameters from input. Context Management Manages state and flow of multi-turn conversations using short-term and long-term context. Session Management - Keeps track of a unique user’s conversation, including past intents and variables. Dialog Management- Determines how to respond based on intent, entities, and context. Fulfilment Module - Can connect to external APIs or backends to perform actions.
-
Fulfilment & Webhooks Fulfilment is an optional but powerful part of Dialogflow that allows dynamic responses. Webhook Integration: Sends a POST request to your server with intent and entity data. Backend Logic: Processes the request Response Returned: Sent back through Dialogflow to the user.
4.1.2 - Phase
Production
4.1.3 - Maintenance
The DVLA chatbot team undertake observations to monitor the accuracy of the information it is providing.
There is an established resolution path which routes to the Digital Service Support Team to address these in case of issues.
Google AI Dialogflow Maintenance & Update Frequency 1. Continuous Backend Updates Google frequently updates Dialogflow’s backend services, including: NLU models (improving intent detection and entity extraction) Stability, speed, and scalability enhancements Bug fixes and security patches These updates are automatic and do not require DVLA intervention.
-
Scheduled Platform Maintenance Planned maintenance windows (usually for infrastructure or Cloud dependencies) are announced in advance via: Google Cloud Status Dashboard Google Cloud Console notifications Maintenance is typically non-disruptive, and high-availability strategies are in place.
-
Feature Releases Dialogflow has regular feature releases, especially in Dialogflow CX: New API endpoints Expanded language support Better integration tools (e.g., for Contact Center AI, Cloud Functions)
These updates are usually announced in: Dialogflow release notes Google Cloud blog or changelog
4.1.4 - Models
Natural language understanding (NLU)
Tier 2 - Model Specification
4.2.1 - Model name
Google AI DialogFlow
4.2.2 - Model version
gemini‑2.0‑flash‑001 & gemini‑2.0‑flash‑lite‑001
4.2.3 - Model task
To understand user intent from natural language inputs, recognise what the user wants to achieve, extract relevant information from user queries, identify key parameters, manage the flow of conversation across multiple turns, reply with predefined messages
4.2.4 - Model input
Text-based customers questions in English
4.2.5 - Model output
To provide users with predefined answers to their questions or to obtain the relevant customer query information so that it may be passed onto the customer service agent.
4.2.6 - Model architecture
The detailed internal model architecture of Google’s Gemini-2.0-flash isn’t publicly disclosed in full by Google, but this is what is inferred from information known.
-
Foundation: Transformer-based Large Language Model (LLM) Gemini-2.0-flash is a Transformer-based LLM Likely employs stacked transformer encoder-decoder layers optimised for both understanding and generation.
-
Flash Optimisation FlashAttention is a fast, memory-efficient attention algorithm improving inference speed and reducing GPU memory usage.
-
Pretraining & Fine-tuning Pretrained on massive multilingual datasets (text, code, dialogue data). Fine-tuned specifically for: Dialogflow conversational tasks Reasoning and contextual understanding Instruction following
-
Parameter Scale While exact size is not public, Gemini-2.0 models are multi-billion to possibly over 100 billion parameters scale, putting them in the class of large foundation models.
4.2.7 - Model performance
Google’s Intent Detection Accuracy Dialogflow’s intent classification models are designed for high accuracy in practical application: 90%+ accuracy or higher on well-trained intents with good example phrases. Accuracy depends heavily on quality and quantity of training data provided by developers.
Google’s transfer learning and pretraining on large language datasets help boost generalisation.
Entity Extraction Precision Prebuilt system entities have very high precision and recall due to extensive pretraining and engineering.
Dialogflow is optimised for low latency suitable for real-time conversations. Google Cloud infrastructure allows scaling to millions of queries per day with consistent performance.
4.2.8 - Datasets
The chat bot did not require training, it uses NLP to understand customer intent and routes to the pre-defined scripts and flows built by the Contact Centre Administration Team
4.2.9 - Dataset purposes
N/A
Tier 2 - Data Specification
4.3.1 - Source data name
Internally developed chatbot scripts and flows
4.3.2 - Data modality
Text
4.3.3 - Data description
General information around Vehicle registration, Vehicle Tax, Private Registrations and Driving Licences that is available on GOV.UK or the internal DVLA Knowledgebase
4.3.4 - Data quantities
DVLA chat bots consist of 158 individual flows with 1584 individual answers.
4.3.5 - Sensitive attributes
Personal information is collected by the Chatbot. Information collected is: Name Address D.O.B Vehicle Registration Number Driving Licence Number
4.3.6 - Data completeness and representativeness
If the customer does not provide the data they will still be routed through to an advisor. There is no integration between the chat bot and DVLA records. It is the advisors responsibility to confirm these details with the customer in line with our data security guidelines
4.3.7 - Source data URL
N/A
4.3.8 - Data collection
When a user interacts with the Chatbot, their conversation is collected and stored according to DVLA Policy. The transcript of the chat (both advisor side and chat bot side) is stored in ‘Storm Recorder’ (our call recording and transcript recording product) for 90 days and is then deleted. The information is only collated in the chat bot and web chat interface to respond to the customer’s enquiry
4.3.9 - Data cleaning
N/A
4.3.10 - Data sharing agreements
N/A not required
4.3.11 - Data access and storage
DVLA Personnel have access to the platform. DVLA is responsible for the data collected and processed as it is the controller. This tool is hosted on UK cloud tenant.
Access to Google DialogFlow is limited to 8 Telecoms Developers and a team of 4 Contact Centre Chat Bot administrators.
Access to Storm Recorder is licence based and is only provided to line manager and Quality Team members whose role requires them to have access to the transcripts for quality and evaluation purposes
Tier 2 - Risks, Mitigations and Impact Assessments
5.1 - Impact assessment
A Data Protection Impact Assessment (DPIA) has been completed and will be updated if any changes to the service are undertaken.
The most current iteration was updated in April 2025.
5.2 - Risks and mitigations
The main risk linked to the use of chatbot was the chatbot providing the customer with incorrect information relating to their query. The main mitigation is that the chatbot will transfer a customer to an advisor if the bot does not understand the enquiry or upon request by the customer. The chatbot does not make any significant decisions. Additionally the Contact Centre Administration Team continually maintains the chat bot and ensures the information provided is the latest and most accurate advice