Back to portfolio

Service design project

Customer Enquiry
AI Workflow

Service Design

I designed an AI-powered enquiry system that captures, understands and resolves customer questions faster — across multiple channels while giving agents the context they need.

Role
Service Designer & UX Strategist
Timeline
Continuous Life Cycle
Focus
End-to-end service design, AI workflow, CX improvement
Channels
Website, Live Chat, Email, WhatsApp, Social Media
AI customer enquiry dashboard shown on a laptop

Impact at a glance

-42%

Reduction in response time

vs. before

+35%

First contact resolution rate

vs. before

+28%

Customer satisfaction (CSAT)

vs. before

-30%

Agent handling time

vs. before

£21.6k

Annual operational savings

estimated

About the project

Customers contacted us across multiple channels asking similar questions. Responses were slow and inconsistent, agents lacked context, and valuable time was lost on repetitive queries.

We needed a smarter, faster and more consistent way to help our customers.

The challenge

  • Enquiries coming in across multiple channels
  • Similar questions answered repeatedly
  • Slow and inconsistent responses
  • Agents lacked context and history
  • High handling time and low first contact resolution
  • No unified view of customer enquiries

Objectives

  • Unify enquiries from all channels in one intelligent inbox
  • Use AI to understand intent and suggest best responses
  • Automate answers for common questions
  • Escalate complex issues to the right agent
  • Improve customer satisfaction and reduce response time
  • Create insights to continuously improve the service
Better conversations. Happier customers. Empowered agents.Designed for people. Powered by AI.

Service blueprint

End-to-End Service Blueprint

Mapping the customer journey and backstage processes helped us design an AI workflow that delivers fast, accurate and contextual support.

1. Enquire(Channels)
Customer Journey
Customer asks a question
Frontstage
Web, Live Chat, Email, WhatsApp, Social Media
Backstage
Enquiry logged and channel normalised
Technology
Omnichannel Gateway
Support
Customer
2. Understand(AI Intake)
Customer Journey
Enquiry captured and intent identified
Frontstage
AI intake captures enquiry and extracts intent & entities
Backstage
NLP classifies intent and checks knowledge base
Technology
NLP / Intent Detection
Support
AI System
3. Resolve / Assist(AI + Agent)
Customer Journey
AI suggests answer or routes to agent
Frontstage
AI answers instantly or agent handles complex enquiry with context
Backstage
Workflow routes or escalates with full context
Technology
AI Engine + Agent Workspace
Support
Support Agents
4. Respond(Customer)
Customer Journey
Customer receives accurate response
Frontstage
Response delivered in channel of choice
Backstage
Response tracked and conversation closed
Technology
Response Composer
Support
Customer
5. Learn & Improve(Insights)
Customer Journey
Feedback captured and knowledge improves
Frontstage
Customer rates experience (Optional)
Backstage
Insights analysed and content updated
Technology
Analytics & Knowledge Management
Support
CX & Knowledge Team

AI workflow in action

  1. Enquiry Received

  2. AI Understands Intent

  3. AI Answers or Suggests

  4. Agent Handles (if needed)

  5. Resolved & Learned

AI workflow overview

All channels feed into a single intelligent inbox.

AI identifies intent, extracts key details and checks knowledge.

AI provides instant answer or suggests best response.

Agent reviews with full context and responds.

Outcome tracked, feedback captured, system learns.

Key design decisions

  • Human-centred design

    Ensured the AI supports agents, not replaces them.

  • Intent-first approach

    Improves accuracy and reduces friction.

  • Clear escalation paths

    Ensure complex issues get human expertise.

  • Continuous learning loop

    Uses feedback to make the system better daily.

Tools & technologies

  • OpenAI
  • Make
  • Freshdesk
  • HubSpot
  • Dialogflow
  • Zapier

The AI handles the routine, so we can focus on what really matters, solving the complex problems.

Support Team LeadInternal feedback

Real impact

(From live analytics)

1,566

AI Bot Interactions

YTD

135

Unique Pages Accessed

YTD

21,817

Total AI Bot Hits

This Year

984

Unique Pages

This Year

What's next

  • Extend AI capabilities for proactive customer notifications
  • Add voice (IVR) AI for phone enquiries
  • Expand self-service knowledge base with AI content generation
  • Predict issues and offer proactive help
  • Integrate customer sentiment analysis

Future opportunities identified for the next phase — not yet released.

Key takeaway

By combining service design with AI, I created a system that delivers faster answers, happier customers, and more empowered teams.