How an engagement actually runs.
No mystery. Here's the timeline, what we'll need from you, how we measure success, and how often we'll talk.
Engagement timeline
A typical Launch or Accelerate engagement — adjusted for your scope.
- Week 0
Discovery & scoping
30-min intro, scoped quote within 24 hours, and a kickoff call. We confirm top intents, target deflection, success metrics, and access plan.
- Weeks 1–2
Workflow & action design
We design the intents, the agent flows, and the actions in your real systems. You review and approve. We start integrating.
- Weeks 2–4
Build & integrate
Hands-on configuration in your AI platform, integration with your help desk and internal APIs, knowledge base structuring, and staging.
- Weeks 4–6
Launch
Phased rollout to production. We monitor deflection, intent coverage, and CSAT closely and tune in real time.
- Ongoing
Optimize
Weekly tuning, intent expansion, and action coverage growth. Deflection compounds over the first 90 days.
What we need from you
We do the heavy lifting. We just need a small amount of access and one engaged owner.
- A point person on your side (CX ops, support eng, or PM)
- Access to your help desk and AI platform (we can sit under your SSO)
- Top intents + a sample of tagged tickets (we'll help if untagged)
- Read/write API access for the actions you want the agent to take
- Stakeholder availability for weekly 30-min reviews during build
How we measure success
Every engagement is scoped against numbers that show up in your weekly business review.
Deflection rate
Automated resolutions ÷ total eligible tickets. The headline metric.
Time to launch
Days from kickoff to first production traffic.
Resolution quality
CSAT on automated resolutions vs. human-handled baseline.
Action coverage
Share of top intents the agent can fully resolve without handoff.
Communication cadence
Weekly review (30 min)
Deflection, intent performance, what's next.
Async on Slack
Shared channel for fast questions and updates.
Quarterly business review
On Partner engagements — roadmap and big-picture metrics.