If you use (or are going to use) Kommo CRM with WhatsApp and wants to sell more without hiring more people, the path is simple: design SLA, queues, and routing as if it were an operation. It's not about “setting up a pretty funnel.” It's about implementing clear rules of speed, distribution, and priority — and measuring this every day.
I am Luiz Otávio Gonçalves. A Mechanical Engineer who became a pioneer in Digital Engineering in Brazil. I implement Kommo in real operations, with ROI accountability and with the numbers closing. And I'll be direct: if you are not willing to follow method and data, this post is not for you. If you want to stop losing leads on WhatsApp, here is the design.
The angle almost no one addresses: WhatsApp is a queue (not a chat)
The number 1 mistake I see: treating WhatsApp as a “conversation” and not as a queue system. In high-volume operations, WhatsApp becomes:
- Incoming queue (new leads + returns + off-hours messages)
- Priority queue (hot vs warm vs support vs post-sale)
- Owner queue (each conversation needs a “who is responsible”)
- SLA queue (maximum time for first response and maximum time until proposal)
Without this, what happens is predictable: two agents respond to the same lead, no one takes ownership of the follow-up, the client disappears, and you conclude “the lead is bad.” It's not. Your operation is out of control.
Before talking about Kommo: define 4 SLAs that unlock revenue
SLA is not bureaucracy. SLA is a revenue mechanism. I define these 4 (minimum):
- SLA 1 — First response: maximum time to respond to a new lead (e.g., 5 minutes during business hours).
- SLA 2 — Qualification: maximum time to collect minimum data (e.g., up to 20 minutes after first response).
- SLA 3 — Next step: maximum time to schedule/send proposal (e.g., up to 2 hours).
- SLA 4 — Follow-up: cadence and deadline for follow-up (e.g., Day+1, Day+3, Day+7).
Honesty: if your ticket is low and your volume is small, you can simplify. But if you generate paid leads and have no SLA, you are burning CAC.
If you want to deepen how to measure automation results, I've already shared the method here: how to measure the ROI of business automations.
What to implement in Kommo: 3 layers (funnel, queues, and automations)
In Kommo, I like to think in 3 layers. If you skip one, you “have CRM,” but you don't have a scalable operation.
Layer 1 — Funnel that reflects decision (not pretty stages)
Your funnel is not an organizational chart. It is a decision model. For WhatsApp, the basic that works (in most niches) is:
- New lead (no contact)
- In service (qualifying)
- Proposal/Quote sent
- Negotiation
- Won / Lost
- Post-sale / Retention (if there is repurchase)
I have already explained funnel traps by niche elsewhere on the blog. If your CRM “doesn't move,” this is where it starts to stall: CRM by niche: the number 1 mistake that destroys ROI.
Layer 2 — Queues and routing (the part that generates SLA)
This is where ROI appears. You need clear routing. I use this map:
- Queue A — New leads: everything that comes from ads, organic, landing page, and direct.
- Queue B — Return/Follow-up: leads who have already spoken and need follow-up.
- Queue C — Proposal/Budget: the lead is “hanging” waiting for something from you.
- Queue D — Post-sale/Retention: NPS, support, repurchase.
Golden rule: if the same agent mixes everything in the same “bucket,” priority becomes emotion. And emotion doesn't scale.
Layer 3 — Automation that eliminates micro-decisions
Good automation is the one that eliminates repetitive micro-decisions from the agent. Practical examples:
- Upon entering “New lead,” create task “Respond within X minutes” and notify in internal channel.
- If there is no lead response within 24h, move to “Follow-up D+1” and send a short message (no long text).
- If the lead says “I want to schedule,” automatically send time options and create a task for a human to confirm.
- If the lead asks for price without context, trigger a minimum qualification question (without turning it into an interrogation).
Want to build automations from scratch with method? I showed the way here: how to create automation flows from scratch.
How much it really costs to get Kommo + WhatsApp running (without guessing)
I will separate cost into 3 blocks, because this is where most get it wrong:
| Block | What is | How it impacts ROI |
|---|---|---|
| Kommo license | Monthly fee per user (varies by plan) | Defines fixed cost per seat and predictability |
| WhatsApp (API) | Provider + charge per conversation (WhatsApp Business Platform model) | Defines variable cost by volume and quality of service |
| Implementation | Project: funnel, integrations, automations, training, governance | Defines speed until ROI (and whether you will stall) |
Real fact (without making up numbers): WhatsApp in API mode (WhatsApp Business Platform) is charged per conversation (official Meta/WhatsApp model). This is not opinion; it is the product model. The price in local currency varies by country, conversation category, and policy changes over time, so I won't guess values here.
If you want a complete view of Kommo plans and prices, I detailed it in another post: How much does Kommo CRM cost? Prices and plans. And if you want to understand the real cost per conversation on WhatsApp (where the cash flow bleeds), I recommend this one: Kommo CRM with WhatsApp: how to calculate the real cost per conversation.
Implementation by “operation scenario”: what changes in SLA and routing
Now the important part: how I design when the company grows — without turning Kommo into an expensive spreadsheet.
Scenario 1 — Small team (1 to 3 agents): focus on discipline, not complexity
- One funnel and at most 2 queues: New and Follow-up.
- A clear SLA (first response) and a simple cadence.
- Forbidden: 20 automations and 50 tags. You will only get confused and stop.
Suitable for: clinics, consulting, local services, reasonable average order value, and low/medium volume.
Not suitable for: operations with 300 leads/day on WhatsApp. There you need flow engineering.
Scenario 2 — Medium team (4 to 15 agents): routing + daily audit
- Routing by source (paid traffic, referral, organic) or by product (line A/B/C).
- Daily audit: every day someone checks “leads without owner”, “leads without response”, “overdue proposal”.
- Proposal queue mandatory. Those who don’t create this keep losing sales in the “I’ll check and get back to you” phase.
This is the point where I see the most ROI: you stop wasting leads that were already warm.
Scenario 3 — Large operation (16+ agents): separate roles and lock governance
- SDR/Pre-sales separated from Closer/Sales (even if it’s “part-time”).
- Routing by capacity (whoever is free takes new) + exception rules (VIP, recurrence, churn).
- SLA dashboard + automatic alerts (if SLA is exceeded, priority rises).
- Governance: who can create stages, who can create automation, field standards.
If you’re still deciding which automation tool to use to integrate everything (Kommo, WhatsApp, spreadsheets, ERP, etc.), I left a very pragmatic guide: how to choose the ideal automation tool.
Practical checklist: how I build SLA and queues in Kommo (step by step)
I don’t improvise. I execute this checklist:
- 1) Map inputs: where conversations come from (WhatsApp, Instagram, website, referral).
- 2) Define lead owner: which rule assigns responsibility (round robin, by team, by product).
- 3) Create measurable SLAs: minutes and hours, not “as soon as possible”.
- 4) Create automatic tasks: a task is the unit of execution. Without a task, it becomes “memory”.
- 5) Create exception states: out of hours, waiting for document, waiting for payment.
- 6) Define cadence: minimum viable follow-up (D+1, D+3, D+7) with short messages.
- 7) Implement audit: one person responsible for checking the funnel every day (15 minutes).
- 8) Measure 3 numbers: first response time, stage advancement rate, and conversion by source.
Who this is for (and who it’s NOT for)
Serves If you:
- Sells via WhatsApp and loses leads due to delay, lack of owner, or lack of follow-up.
- Already invests in traffic/referral and wants to optimize CAC without hiring more.
- Willing to operate with process: SLA, queue, audit, and continuous improvement.
Is not useful If you:
- Wants “a CRM that solves itself” without a team responding with discipline.
- Does not have a minimum lead volume and will struggle to justify the project (first fix offer and acquisition).
- Will use unofficial WhatsApp integration and then blame the tool when banned/operation stalls.
The result I seek: more conversations turning into proposals, more proposals turning into revenue
When SLA and queues are implemented, you quickly see:
- Drop of “forgotten” leads (without owner/without response)
- More speed in the first response (which impacts conversion)
- More consistency in the follow-up (where most sales happen)
- Less stress in the team (because priority stops being a shout)
Want me to design this in your operation (Kommo + WhatsApp + automations + governance), focusing on ROI and execution? Then the next step is objective: request a project.
FAQ
The questions below are the ones I answer most when the topic is Kommo + WhatsApp with real operation.
- Does Kommo replace WhatsApp Business?
It depends on your model. In volume operations, Kommo becomes the “center” of customer service and WhatsApp becomes a channel. The critical point is to use official integration (API) when you need scale and governance. - Do I really need SLA if I have few leads?
If you have few leads, a simple SLA solves it (e.g., respond within 10 minutes). What you can’t do is operate on a “when possible” basis. That kills conversion. - How to avoid two agents responding to the same lead?
With a “lead owner” rule + new queue + automatic distribution. Without an owner, it becomes no man’s land. - What is the most important metric on WhatsApp?
To start: First response time and Advancement rate stage progression (New → Service → Proposal). Without this, you argue opinions. - Does automation on WhatsApp make customer service robotic?
It does if you automate what is human (negotiation, complex objection). I automate screening, SLA, tasks, and basic follow-up — and leave the human where they really close.