Developer Handover Pack

Developer Handover Pack

Proposal & SOW Creation

Reads the scoping notes recorded against a HubSpot deal and extracts modules, assumptions, constraints, and any open questions into a structured JSON scope object. The output is stored back on the HubSpot deal record and passed to the Estimate & Proposal Agent. No LLM hallucination risk on pricing at this stage as no numbers are produced here.

// Input
HubSpot deal ID + associated note body text
deal.properties.hs_latest_meeting_activity (note body)
deal.properties.dealname

// Processing
LLM extraction prompt -> structured scope object
Keys: modules[], assumptions[], constraints[], open_questions[]

// Output
deal.properties.scope_json = { modules, assumptions, constraints, open_questions }
deal.properties.scope_status = 'ready' | 'incomplete'
Trigger: POST /estimate-agent with deal_id and scope_json

Receives the structured scope object, applies the stored rate card to each module, produces a line-item effort estimate, and then generates a full proposal document and matching SOW using the approved Google Docs templates. Outputs two drafted documents on the deal record. If the calculated total falls outside the pricing range threshold a flag is raised for the Studio Owner review gate before the Send & Track Agent is triggered.

// Input
deal_id + scope_json from Scope Extraction Agent
rate_card config: { module_name, hours_low, hours_high, day_rate }
Google Docs template IDs: proposal_template_id, sow_template_id

// Processing
rate_card.apply(scope_json.modules) -> line_item_estimate[]
total_price = sum(line_item_estimate[].price)
pricing_in_range = total_price <= rate_card.ceiling
Google Docs API: copy template -> fill placeholders -> return doc_url

// Output
deal.properties.proposal_doc_url = '<Google Docs URL>'
deal.properties.sow_doc_url = '<Google Docs URL>'
deal.properties.estimate_total = total_price
deal.properties.pricing_flag = 'in_range' | 'review_required'
Trigger (if in_range): POST /send-track-agent with deal_id
Trigger (if review_required): Notify Studio Owner via HubSpot task

Takes the approved proposal document, uploads it to PandaDoc, assigns signatories, and sends it to the prospect. Polls PandaDoc for document status updates (opened, signed, declined) and writes status back to the HubSpot deal record in real time. Triggers a follow-up task on the deal if the document has not been opened within 48 hours.

// Input
deal_id + proposal_doc_url + contact email from HubSpot
PandaDoc API key (workspace-scoped)
Signatory: deal.contacts[0].email, deal.contacts[0].firstname

// Processing
PandaDoc: POST /documents { file_url: proposal_doc_url, recipients: [contact] }
PandaDoc: POST /documents/{id}/send
PandaDoc webhook -> FullSpec listener -> parse event.type

// Output
deal.properties.pandadoc_document_id = '<PandaDoc doc ID>'
deal.properties.proposal_status = 'sent' | 'viewed' | 'signed' | 'declined'
deal.properties.proposal_sent_at = ISO8601 timestamp
HubSpot Task (if no open after 48 hrs): 'Follow up on proposal: {dealname}'

// ============================================================
// TRIGGER
// ============================================================
HubSpot deal.stage -> 'Scoping Call Done'
  Webhook: POST https://app.fullspec.io/hooks/inbound/proposal-sow
  Payload: { deal_id, dealname, stage, timestamp }

// ============================================================
// STAGE 1: Scope Extraction Agent
// ============================================================
FullSpec receives webhook
  -> HubSpot GET /deals/{deal_id}/notes
     response: { note_body: '<raw scoping notes text>' }
  -> LLM extraction prompt:
     input:  note_body
     output: scope_json {
       modules:      ['Module A', 'Module B', ...],
       assumptions:  ['Assumption 1', ...],
       constraints:  ['Timeline: 6 weeks', ...],
       open_questions: ['Q1', ...]
     }
  -> Validate: all keys present AND modules.length > 0
     if INVALID -> HubSpot task: 'Scope extraction incomplete, manual review needed'
     if VALID   -> HubSpot PATCH /deals/{deal_id} { scope_json, scope_status: 'ready' }

// ============================================================
// STAGE 2: Estimate & Proposal Agent
// ============================================================
Trigger: scope_status = 'ready' on deal record
  -> Load rate_card config (versioned JSON)
     rate_card: [{ module, hours_low, hours_high, day_rate }, ...]
  -> Apply rate card to scope_json.modules
     line_items = modules.map(m => rate_card.match(m))
     total_price = sum(line_items[].price)
  -> Pricing gate:
     if total_price <= rate_card.ceiling
        pricing_flag = 'in_range'
     else
        pricing_flag = 'review_required'
        -> HubSpot Task: 'Review pricing for {dealname}'
        -> PAUSE until deal.properties.pricing_approval = 'approved'
  -> Google Docs API:
     POST /documents/{proposal_template_id}/copy -> new_proposal_doc_id
     POST /documents/{sow_template_id}/copy     -> new_sow_doc_id
     Batch update: fill named placeholders with scope_json + line_items
  -> HubSpot PATCH /deals/{deal_id} {
       proposal_doc_url, sow_doc_url,
       estimate_total, pricing_flag
     }

// ============================================================
// STAGE 3: Human Review Gate (retained by design)
// ============================================================
Studio Owner reviews proposal_doc_url + sow_doc_url
  Sets: deal.properties.pricing_approval = 'approved' | 'rejected'
  if 'rejected' -> process terminates, task created for Account Lead
  if 'approved' -> pipeline continues to Send & Track Agent

// ============================================================
// STAGE 4: Send & Track Agent
// ============================================================
Trigger: pricing_approval = 'approved' OR pricing_flag = 'in_range'
  -> HubSpot GET /deals/{deal_id}/contacts
     contact: { email, firstname, lastname }
  -> PandaDoc POST /documents {
       file_url: proposal_doc_url,
       recipients: [{ email, firstname, lastname, role: 'signer' }]
     }
     response: { pandadoc_document_id }
  -> PandaDoc POST /documents/{pandadoc_document_id}/send
  -> HubSpot PATCH /deals/{deal_id} {
       pandadoc_document_id,
       proposal_status: 'sent',
       proposal_sent_at: ISO8601
     }

// ============================================================
// STAGE 5: Status Tracking (PandaDoc -> HubSpot sync)
// ============================================================
PandaDoc webhook -> POST https://app.fullspec.io/hooks/pandadoc/status
  event.type = 'document.viewed'   -> deal.proposal_status = 'viewed'
  event.type = 'document.completed'-> deal.proposal_status = 'signed'
  event.type = 'document.declined' -> deal.proposal_status = 'declined'
                                   -> HubSpot Task: 'Proposal declined: {dealname}'
48-hour check (scheduled job):
  if proposal_status = 'sent' AND now - proposal_sent_at > 172800s
     -> HubSpot Task: 'Follow up on proposal: {dealname}'

// ============================================================
// TERMINAL OUTPUT
// ============================================================
deal.stage -> 'Proposal Sent'
deal.properties: { proposal_status, pandadoc_document_id, estimate_total, proposal_sent_at }