What happens, week by week
Every milestone is defined before we start. You always know where we are and what's coming next.
Discovery & Assessment
assessKnowledge Preparation
preparePersonality & Brand Build
prepareBuild, Test, Refine
buildIntegration & Launch
launchLive Monitoring & Optimisation
liveWhat we need from you
We've designed this process to require as little of your time as possible.
Your existing documents
FAQs, product descriptions, pricing, policies, process guides — whatever you have. We work with what exists.
Common customer questions
A dump of your inbox, a list off the top of your head, or a CRM export. Any format works.
10 hours of your time
Spread across the 4-week build: one discovery session, one review session, and light feedback on test results.
One point of contact
We don't need a committee. One person with authority to approve the agent's personality, tone, accuracy and give regular feedback.
Most clients spend fewer than 10 hours total throughout the entire 4-week build.
Honest answers to honest questions
Questions we hear regularly during discovery calls — answered directly.
We've tried chatbots before and they didn't work.
Most chatbot implementations fail because of poor knowledge preparation. The agent answers from whatever you feed it. If that's a pile of unstructured documents, the answers will be inconsistent. Our entire process is designed around fixing this: knowledge preparation is weeks 1 and 2, not an afterthought.
We don't have technical staff. Can we still do this?
Yes — completely. We handle all technical work. For your website, integration is a single embed snippet that takes minutes to add. If you work with a web agency, we coordinate with them directly. You don't need to understand anything technical to benefit from this.
What if our knowledge base changes after launch?
Knowledge bases always evolve. We provide you with access to an included bonus 'Your Brain Blueprint' so you can identify what needs updating. On the maintenance plan, we handle updates as part of the monthly optimisation cycle.
How do you ensure the agent doesn't go off-script?
Behavioural guardrails are built in during the construction phase — hard limits on what topics the agent engages with, how it handles sensitive or inappropriate requests, and when it escalates to a human. This is tested extensively before launch.