Skip to content

Roadmap

Shane Rounce edited this page May 12, 2026 · 5 revisions

Roadmap

I am building this in deliberate phases so the plugin stays coherent.

Phase 1: first-party data spine

This is the phase already in place:

  • sessions and events
  • number routing
  • admin reporting
  • privacy controls
  • enrichment groundwork
  • call schema and reporting surfaces
  • sample data

Phase 2: deeper enrichment and reporting polish

This is where I want to keep improving:

  • drilldown quality
  • ranking and chart usefulness
  • reporting links and follow-up workflow
  • broader enrichment polish and interpretation

Phase 3: Amazon Connect integration

This will cover:

  • broader call import and matching flows
  • stronger call matching
  • deeper number and contact-flow alignment
  • later outbound callback support

The setup screens already prepare for this, and the current build now includes readiness checks, live claimed-number visibility, available-number search, in-plugin number claiming, live flow tools, and an operational call-import path plus import-status view.

The larger call-import and matching layer is still a key next step. I do not want to blur that with the separate website-assistant work now that the frontend chat direction has moved back into WordPress itself.

Phase 4: AI-assisted engagement

This is now partly in motion rather than purely future work.

What is already in place:

  • a WordPress-managed OpenAI assistant for company, product, and service questions
  • live site-context grounding from published WordPress and WooCommerce content
  • stored frontend chat conversations and reporting
  • richer frontend source presentation with inline links, thumbnails, and collapsed extra matches
  • live operator handover, typing state, and admin reply suggestions
  • passive lead capture and lightweight repeat-visitor company/contact memory
  • WooCommerce shipping-aware answers using live zones, basket context, quantity hints, and postcode follow-ups where the site setup supports it

What I still want to keep improving:

  • safer assistant tooling and boundaries
  • stronger answer shaping and citations
  • better control over handoff and escalation behaviour
  • eventually voice-oriented experience where appropriate

The assistant is live enough to be useful now, but I still treat the presentation, safety model, and lead-capture behaviour as active refinement work.

Ongoing principle

I would rather keep each phase complete and dependable than rush speculative features into the codebase before the underlying reporting and setup model is ready.

Clone this wiki locally