Guided service selection
Help people explain their goal, ask purposeful clarification questions and direct them to a suitable service without forcing them to understand internal department names.
AI agents & knowledge
Conversational AI
Design consistent, useful conversations across web, messaging and voice channels.
Built around a specific business need.
A conversation can fail even when each individual answer is factually correct. The assistant may ask for information already provided, lose track of the user’s goal or move them between channels without useful context. Conversational AI design addresses that wider experience: how dialogue progresses towards a worthwhile outcome.
Kent Consultancy would shape the conversation around user needs, channel constraints and service operations. The work can span web chat, messaging and voice, with shared guidance where appropriate and channel-specific interactions where necessary. We would define when context may be carried forward, how identity affects the conversation and how an adviser can take over without making the user start again.
Who this is for
- Organisations operating several customer contact channels with inconsistent experiences.
- Service teams whose current assistant answers questions but struggles with journeys.
- Product owners developing guided, conversational access to business services.
Where this can make a practical difference.
Cross-channel continuity
Design a controlled transfer from one contact channel to another, sharing only the context needed and only where identity, consent and integration arrangements support it.
Complex enquiry navigation
Break a multi-part request into manageable steps, confirm understanding and involve a person where the conversation requires discretion, empathy or specialist judgement.
How we approach the work.
Research real dialogue
Review existing enquiry journeys, common misunderstandings and abandonment points. Identify what users are trying to achieve and how they describe those needs in their own words.
Model conversation state
Define what is known, what still needs clarification and when information becomes stale. Decide which context belongs only to the session and which may be retained.
Design channel behaviour
Write prompts, repair strategies and handover messages for each channel. Spoken interactions need different pacing from text, while messaging may involve long pauses between replies.
Connect service actions
Map dialogue stages to permitted lookups or requests. Verify identity before discussing protected records and confirm consequential choices before passing them to operational systems.
Evaluate whole journeys
Test incomplete replies, changed intentions, repeated questions and requests for a person. Assess whether the user reaches a useful outcome, not simply whether the assistant sounds natural.
The sequence is tailored to the agreed scope, dependencies and risk. It describes our delivery approach, not a fixed timeline or performance guarantee.
What the engagement can deliver.
- User journey and conversation-state designs.
- Channel-specific dialogue and recovery patterns.
- Context-sharing and identity requirements.
- Integrated conversation flows for the agreed channels.
- Journey evaluation findings and an improvement backlog.
The proposal identifies the deliverables, acceptance evidence, responsibilities and support arrangements included in your engagement.
Measure what matters to this workflow.
Journey completion
Users reaching the intended service outcome within the tested scope.
Conversation repair
Misunderstandings resolved without restarting or losing essential context.
Repeated information
How often users must provide details already available appropriately.
These are measures to discuss and baseline. Targets are agreed for the specific system and data; no universal improvement percentage is assumed.
What to prepare for discovery.
- Examples of real enquiries and known customer journey difficulties.
- Access to channel owners and front-line service colleagues.
- Agreed rules for identity, session history and human escalation.
Questions worth asking.
Is conversational AI a chatbot?
A chatbot can be one interface within the solution. This service focuses on dialogue design, context and service continuity, which may extend across several interfaces.
Should every channel behave identically?
The service meaning should remain consistent, but interaction patterns should fit the channel. A long text explanation, for example, usually needs restructuring before it works well in speech.
Can conversations be personalised?
Where there is a clear purpose and appropriate access, selected context can help. The design should make the basis of personalisation understandable and avoid using unnecessary or speculative personal information.