Carisam · Sprint Closeout · Source documents
Back to closeout page

DOCUMENT 05

Next-phase roadmap

What Phase 2 could look like. Reference only — no commitment.

CARISAM Next Phase Roadmap

Executive read

The first sprint identified the right starting point — faster quote turnaround — and delivered the working operating layer in nine days at https://carisam-quote-lane.skippi.ai/. Lance and Jim can capture requests, send vendor RFQ links, compare responses, mark decisions, and export pro-forma CSV today.

The next phase extends from quote-side coverage into order-to-closeout, with the integrations and hardening that turn the deployed app into the durable CARISAM operating system.

What is already live (sprint 1)

These were originally phased as future work; they are in production now:

  1. Internal quote intake — every new request lands in one record, with customer, project, destination, source, priority, line items, notes.
  2. Vendor RFQ response path — generate copy-pasteable RFQ email plus a unique unguessable link per vendor. No vendor account needed.
  3. Vendor response capture — pricing per item, lead time, availability, freight, validity, confidence, notes — all submitted through one form.
  4. Side-by-side comparison view — across all vendor responses for a quote, highlighting lowest price and fastest lead time.
  5. Decision capture — pick winning response, margin %, customer sell total, decision notes.
  6. Pro-forma CSV export — populates vendor totals × margin into sell unit / sell total columns, ready for the CARISAM workbook.

Phase 2 scope (next sprint)

A. Customer quote output — branded PDF

Goal: generate a customer-ready PDF in CARISAM's existing pro-forma layout, with letterhead, terms, signatures.

Functions:

  1. PDF template matching CARISAM brand.
  2. Pulls all data from the live app (no manual retyping).
  3. Editable cover letter / terms section.
  4. Send-ready file Lance can email directly.
  5. Source data still linked for audit.

Success measure: a real customer quote leaves the system as a polished PDF, never re-typed.

B. Order conversion — quote becomes order

Goal: turn accepted quotes into order records with their own tracking surface.

Functions:

  1. Customer PO field.
  2. Vendor PO field.
  3. Deposit received / balance due tracking.
  4. Vendor payment status.
  5. Production / shipping path / booking number.
  6. Customer-side update status (last touched).
  7. Closeout status.

Success measure: active orders can be reviewed without opening multiple spreadsheets and email threads.

C. Email integration

Goal: RFQ goes from the app straight into the vendor's inbox, and vendor email replies show up alongside the structured response link.

Functions:

  1. Outbound RFQ email send (SMTP or Postmark/Resend).
  2. Reply-to capture so email responses attach to the RFQ record.
  3. Optional Gmail/Outlook tab-sync for the existing CARISAM mailbox.

Success measure: Lance never copy-pastes again — RFQs go out from inside the app, responses come back into the app.

D. PDF and email extraction (assisted)

Goal: when a vendor responds by PDF or plain email instead of the link, extract the pricing into the comparison table.

Functions:

  1. PDF parser for common vendor quote layouts.
  2. Email text parser for unstructured replies.
  3. Confidence score per extracted field.
  4. Always Lance/Jim review before use.

Success measure: Lance can drop a vendor PDF into the app and have it pre-fill the response form for review.

E. Multi-user + audit

Goal: Lance and Jim as distinct users, with an audit trail of who did what.

Functions:

  1. Per-user sign-in (replace shared password).
  2. Change log on every quote.
  3. Permissions: Jim can prepare, Lance can approve/send.

Success measure: Jim can run the day-to-day quote prep work, Lance approves and decides, both see the same source of truth.

Steve Dakh — Phase 2 architecture review

Bring Steve in to harden the deployed app before Phase 2 features go on top.

Questions for Steve

  1. Current stack is Python stdlib + SQLite + nginx + systemd on Hetzner. Is that the right base for Phase 2 features, or do we lift to a managed platform?
  2. Email integration: SMTP via Postmark/Resend, or roll our own?
  3. PDF generation: WeasyPrint, Playwright, or external service?
  4. Authentication: roll our own per-user auth, or use a managed provider (e.g. Clerk, Auth0) for Phase 2?
  5. Backup and disaster recovery for the SQLite DB.
  6. Monitoring and uptime alerts.
  7. Multi-device sync considerations as Jim and Lance both work the same quotes.

Recommendation

Stay lightweight. Don't migrate to a heavy CRM. The current app is small, fast, and built around CARISAM's actual workflow. Phase 2 adds the next layer (orders, email, PDF) on the same foundation rather than a rebuild.

The non-negotiables for Phase 2: keep Lance's judgment in the loop, don't let AI extraction be authoritative without review, don't add steps that slow Lance down, preserve the existing CARISAM Excel/PDF format on output.