🔍 Initiative Validation Report — Qontak Automation Flow Builder
Initiative: Extend the bot conversation flow builder into a Qontak-native general automation flow builder (triggers → conditions → actions), centralizing all Qontak automation into one surface
Module: Chatbot / Omnichannel (Qontak)
DRI: Dimas Fauzi Hidayat
Date: 17 Jul 2026 · Desk research via web search, all sources accessed 17 Jul 2026
Companion doc: discovery-product-plan-2026-07-17.md (grounded product plan incl. proposed pricing)
1. Problem Assessment
Workflow automation inside customer-conversation platforms is a mainstream, analyst-backed category, not a speculative bet: the conversational AI market is ~$14–15B (2025) growing ~25% YoY (Research and Markets), and Gartner predicts agentic AI will autonomously resolve 80% of common customer-service issues by 2029 (Gartner, Mar 2025). Indonesia is the #2 WhatsApp Business market globally (~163.5M downloads, ElectroIQ stats), with vendor-relayed figures of ~60% SME WhatsApp API adoption for automation (❓ vendor-sourced, Sobot). The hard economic trigger is Meta's 1 Oct 2026 service-message pricing (every free-form reply becomes billable — corroborated by ≥4 independent BSPs: charles, YCloud, ChakraHQ, Sleekflow), which makes fewer-message, higher-containment automation directly money-saving for every WhatsApp tenant.
Signal Strength: 🟢 Strong — analyst backing + hard market timing + internally verified pain (scattered point automations, tree-diagram performance complaints).
2. Competitor Landscape
2A. Capability existence map (verified against official docs, 17 Jul 2026)
| Capability | Qontak (today) | Sleekflow | Respond.io | WATI | Trengo | Gallabox |
|---|---|---|---|---|---|---|
| Visual automation canvas beyond bot conversation | ❌ | ✅ Advanced Flow Builder | ✅ Workflows | ❌ (rules, no canvas) | ⚠️ AI Journeys (beta) | ⚠️ template-shaped Workflows |
| Trigger: conversation lifecycle (open/close/resolve) | ❌ | ✅ | ✅ (no agent-assigned) | ⚠️ AI-detected only | ✅ via Rules | ❌ |
| Trigger: scheduled / time-based | ❌ | ✅ (specific time + recurring) | ❌ explicitly unsupported | ❌ | ❌ | ⚠️ drip-sequence only |
| Trigger: contact/CRM field events | ❌ | ✅ (+ external CRM records) | ✅ tag/field/lifecycle (no contact-created) | ✅ attributes | ❓ | ⚠️ enrollment criteria |
| Trigger: inbound webhook | ❌ | ✅ | ✅ (Advanced plan) | ⚠️ API-start only | ⚠️ on-demand/manual | ❓ |
| Trigger: e-commerce events | ❌ | ✅ Shopify/VTEX | ❌ (ads triggers instead) | ⚠️ add-on | ✅ order/abandoned-cart | ✅ Shopify (max 3 workflows) |
| Action: HTTP request out | ⚠️ (bot tree + AI agent action) | ⚠️ Premium gate | ✅ (Advanced plan) | ❓ not found | ✅ HTTP actions (beta) | ❓ |
| Action: native CRM deal/ticket objects | ⚠️ (AI agent actions exist; not in a canvas) | ❌ external CRMs only | ❌ external only | ❌ | ❌ | ❌ |
| Condition/branch + delay/wait nodes | ⚠️ (branch in bot tree) | ✅ | ✅ | ⚠️ rule filters | ⚠️ weak (documented gap) | ⚠️ fixed delays |
| Run history / per-execution logs | ❌ | ✅ per-enrollment log tab | ❌ open feature request | ❌ (silent failure complaints) | ❌ | ❌ (weakest area) |
Sources: Sleekflow trigger catalog · Respond.io triggers / steps / analytics feature request · WATI rules triggers / actions · Trengo Journeys (beta) / e-commerce triggers · Gallabox docs
2B. Pricing comparison (automation gate + metering)
| Platform | Automation unlocks at | Metering model | Gotchas (from reviews) |
|---|---|---|---|
| Sleekflow | Pro AI ~$149/mo (min 3 users) | Per-enrollment credits: basic flow 0.5 / advanced 1.0 credit per contact-entry; ~3,000/mo included on Pro; add-on $59 per 1,500 | "Unlimited" marketing vs metered docs; node-count changes the credit rate (enrollment docs) |
| Respond.io | Growth $159/mo | No run metering — seat + MAC ($12–15 per extra 100 MACs); caps: 150 workflows × 100 steps | $79 Starter has NO workflows ("a trap" per reviews); HTTP/webhook needs Advanced $279 (pricing) |
| WATI | Growth $99/mo | Per-automation-trigger quota: 1k/2k/5k per tier; every automated execution = 1 trigger, chained = 2 | Automations silently stop at cap — top G2 complaint; ~20% markup on Meta fees (trigger cost docs) |
| Trengo | Effectively Pro €499/mo (Flowbot); Journeys beta unpriced | Seat + metered conversations (7-day window) + €0.25–0.30 per AI conversation from prepaid wallet | Bill-shock reviews (bills doubled post-2024 migration; overage invoices to €9,500) (Chatarmin analysis) |
| Gallabox | Growth $89/mo (6 users) | Bot-conversation allowance/tier; no per-run pricing found | Weak reporting = #1 G2 complaint; add-ons compound |
| Horizontal refs | HubSpot: workflows = Professional gate, ~300 workflows, runs free (FAQ) | Zapier per-task (utility steps free) · Make per-credit · n8n per full execution (steps free, pricing) | Intercom trend: outcome pricing (Fin $0.99/resolution) |
2C. Competitive read
- Not uncharted, but the bar is low and uneven. Only Sleekflow has the full package (canvas + scheduled + webhook + logs). Respond.io has the most general engine but explicitly no scheduled triggers and no run observability. WATI/Gallabox don't really have a workflow canvas.
- Table stakes (must match): visual trigger→condition→action canvas, conversation-lifecycle triggers, assign/tag/field/HTTP actions, delay + branch, templates gallery.
- Differentiators available to Qontak: (1) native CRM objects in the canvas — no peer has deals/tickets as first-class nodes; all bolt on external CRMs; (2) run history/debugging — market-wide hole except Sleekflow; (3) scheduled triggers — only Sleekflow has them; (4) honest, simple pricing — the category's pricing is actively resented (WATI silent-stop, Trengo wallet, Sleekflow credit complexity).
3. Market Landscape
- Demand & trend: conversational automation growing ~25% CAGR globally; Indonesia is the world's #2 WhatsApp Business market; WhatsApp-first automation is the dominant SMB entry point locally (multi-source, §1).
- Commercializable: yes — automation is the proven upgrade lever in this category (Respond.io's $79→$159 jump is the workflow gate; HubSpot sells Professional on workflows). Peers monetize it both as plan-gate and usage meter.
- Niche vs general: general capability, horizontal across all Qontak verticals (commerce, services, finance) — but hero use cases should be verticalized in GTM (COD confirmation, order follow-up, CSAT escalation).
4. Solution Opportunity (solution space)
| Approach | Assessment |
|---|---|
| A. Generalize the existing tree builder | ❌ Rejected — conversation-bound engine + canvas (code-verified; see discovery-product-plan-2026-07-17.md §4) |
| B. Ride Mekari Workflow | ❌ Rejected — cross-BU dependency on the critical path (avoidability assessment in product plan §5) |
| C. Qontak-native builder on the NodeRegistry engine, then migrate bot conversation onto it | ✅ Chosen — full replacement, not a bolt-on; ~70% of the action engine exists; design language exists (qontak-designer bot-flow prototype); table-stakes reachable in 2 phases |
| D. Templates-only automation (no canvas; curated recipes) | Worth keeping as the activation layer on top of C, not a substitute — Gallabox shows template-only caps out fast |
| E. AI-authored automation ("describe it, we build the flow") | Differentiator for a later phase — market is moving here (Gartner agentic prediction); needs C's rails first |
Table stakes = C's P1 scope. Differentiators = one unified engine + native CRM nodes + run observability + scheduled triggers + (later) E.
5. Best Practices & Benchmarks
- Trigger → Condition → Action grammar with trigger-scoped variables and progressive disclosure (trigger first, then conditions, then steps) — incident.io's workflow engineering writeup.
- Run history with per-step data + replay of failed runs — Zapier's trust baseline (replay docs); auto-disable after repeated errors.
- Explicit loop guardrails — HubSpot hard-blocks re-enrollment loops (opt-in re-enrollment per trigger, FAQ); matches the
triggered_by: automationevent-stamping already designed into P1. - Templates gallery as the activation lever (Zapier, HubSpot) — ship 3–5 hero-use-case templates at launch, not an empty canvas.
- Never silently stop automations at quota — WATI's defining anti-pattern; degrade gracefully + notify.
- Utility steps shouldn't cost money — Zapier now excludes filter/delay/format steps from task counts; n8n prices the full run. Simplicity wins trust.
6. Discovery Readiness Verdict
🟢 GO
Problem signal is strong (analyst-backed category, hard Oct-2026 economic trigger, verified internal fragmentation + canvas performance pain), feasibility is exceptional (~70% of the engine exists, design language exists as a Pixel prototype), and the competitive read shows a reachable bar with concrete differentiators (unified engine, native CRM nodes, observability, honest pricing). The main risks are scope gravity ("centralize everything"), pricing/packaging, and the live-traffic migration in P2 — all discovery-addressable, none a reason to pause.
7. Recommended Discovery Method & Questions
Method mix:
- Concept test (primary) — extend the qontak-designer bot-flow prototype with the unified trigger picker + a continuous conversation→automation flow (post-resolve follow-up use case) and test with 6–8 tenant admins/supervisors across segments.
- CS/AM signal sweep — top "I wish it could automatically…" requests from the support/AM ticket base (cheap, fast, validates hero use cases).
- Pricing sensitivity check — lightweight Van Westendorp on the automation add-on/tier framing + run-quota sufficiency per org size.
Priority research questions:
- Which 3 automations would you build in your first week? (validates hero use cases + template gallery content)
- Can non-technical admins parse the trigger→condition→action grammar on the prototype, or do they start from templates? (activation design)
- Where does automation sit in willingness-to-pay — upgrade driver to a higher plan, or paid add-on? (packaging decision)
- What monthly run volume do typical org sizes actually need? (quota calibration; avoid the WATI trap)
- How much do run history/debugging and failure notifications matter vs "set and forget"? (validates the observability differentiator)
8. Commercial Impact
Direct revenue lever — high confidence. Mechanisms: (1) plan upgrades — automation is the proven tier-jump driver in this exact market (Respond.io gates workflows at +$80/mo over Starter; HubSpot sells Professional on workflows); (2) retention/stickiness — tenants with active automations embed Qontak in their operations (n8n/Zapier retention logic); (3) add-on/usage revenue — run packs above included quotas (proposed model in the discovery/product plan §8); (4) defensive — Sleekflow already sells this against Qontak-class platforms in SEA; parity+differentiation protects the base ahead of Oct-2026 WhatsApp cost pressure, when tenants will actively shop for automation that reduces message spend. Indirect: automation events feed the AI-agent action roadmap (same node catalog), compounding the platform story.
Freshness note: competitor features and pricing verified 17 Jul 2026 — re-verify before PRD sign-off and any stakeholder deck. Items marked ❓ in the research could not be confirmed against primary sources.