> whatsapp.agents.workflow.delivery
Never Miss A WhatsApp Lead Again. Turn Inquiries Into Revenue 24/7.
Most businesses lose customers because inquiries wait hours for a reply. We deploy autonomous WhatsApp AI assistants, payment-ready booking engines, and n8n workflow automation that respond in seconds, qualify buyers, and hand off seamlessly to your team.
> starter assistant live in 14 days · fixed pricing in naira · zero lock-in
> fixed.scope.pricing
Turnkey Delivery Packages. Fixed Prices in Naira.
Clear deliverables, rapid timelines, and zero surprises. Tested in production on WhatsApp and integrated directly into your operations.
> whatsapp.starter
Live in 14 DaysStarter WhatsApp AI Assistant
₦350,000
₦65,000 / month
Deploy a dedicated WhatsApp assistant that answers customer questions instantly, captures qualified leads, and hands off complex inquiries to your staff.
- ✓ 24/7 instant replies to FAQs, catalogs, and service pricing
- ✓ Lead qualification & contact capture (Name, Need, Budget)
- ✓ Real-time lead push to Google Sheets, CRM, or email alerts
- ✓ One-tap human handoff with full conversation history
- ✓ Official Meta WhatsApp Business Cloud API onboarding
- ✓ 30-day post-launch prompt tuning & optimization included
> whatsapp.commerce
Live in 21 DaysCommerce & Workflow Engine
₦750,000
₦120,000 / month
Full conversational commerce and appointment scheduling. Process payments, issue receipts, and sync inventory or booking calendars automatically.
- ✓ Everything in Starter Assistant plus:
- ✓ Direct Paystack & Flutterwave checkout links & payment receipts
- ✓ Automated inspection & appointment booking with Google Calendar
- ✓ n8n / Make integration to invoices, accounting, or inventory
- ✓ Automated WhatsApp abandoned follow-ups & reminder sequences
- ✓ NDPA 2023 compliant data isolation and audit logging
> enterprise.workflow
3–4 WeeksCustom Enterprise Workflow
From ₦1,800,000
Custom SLA
High-volume operational automation across multiple internal departments, legacy databases, ERPs, and custom microservices.
- ✓ Self-hosted n8n cluster or custom queue-worker microservices
- ✓ Multi-system reconciliation (Banks, Payment Gateways, ERPs)
- ✓ Internal approval gates on Slack, WhatsApp, and Web portals
- ✓ Multi-agent tool routing with fallback guarantees
- ✓ Dedicated uptime monitoring and priority SLA support
> local.reality
Transparent On API Fees, Naira Costs & Data.
> meta.pricing.transparency
Meta WhatsApp Cloud API Costs
Meta charges per 24-hour conversation category (Service, Utility, Marketing). We structure conversation flows so customer service inquiries fit within free and lowest-cost tiers, minimizing variable Meta invoices.
> predictable.billing
Naira Stability & Caching
We implement prompt caching and semantic routing so repeat inquiries (pricing, addresses, banking details) consume zero external tokens, shielding your business from exchange-rate fluctuations.
> customer.owned.assets
Zero Vendor Lock-In
Your phone numbers remain registered to your own Meta Business Manager. Your workflows belong to your infrastructure. If you stop working with us, your bot remains online.
> the.problem
Most Automation Projects Fail Quietly.
The demo was the product
It worked once, on a clean input, in front of an audience. Nobody scoped what happens on the malformed message at 3am.
Nobody owns the runs
A flow breaks in week six, fails silently, and the team is back to manual work without ever deciding to be.
Per-task pricing bites
The platform bill scales linearly with success. What was clever at 500 runs a month is indefensible at 500,000.
> the.stack
Six Layers of an Automation That Survives.
> 01
Interfaces
The place a human actually meets the system. Most teams do not need another dashboard - they need the workflow to live where conversation already happens.
- ✓ WhatsApp Business API
- ✓ Telegram & Slack apps
- ✓ Web chat & voice
- ✓ Email and forms
> 02
Orchestration
The engine that decides what runs, in what order, and what happens when a step fails. Visual builders where speed wins, code where correctness wins.
- ✓ n8n (self-hosted)
- ✓ Make & Zapier
- ✓ Temporal / queue workers
- ✓ Cron and event triggers
> 03
Reasoning
Models used deliberately: routing intent, extracting structure, drafting language. Deterministic code handles anything a model should never be trusted to do twice.
- ✓ Intent routing & extraction
- ✓ Tool-calling agents
- ✓ RAG over your documents
- ✓ Deterministic fallbacks
> 04
Systems of record
Automation is only as good as the data it writes to. We connect the tools you already run, and build the schema when a spreadsheet has outgrown itself.
- ✓ Sheets & Airtable
- ✓ PostgreSQL & warehouses
- ✓ CRMs, helpdesks, billing
- ✓ Internal and partner APIs
> 05
Guardrails
The difference between a demo and production. Every agent gets limits, approvals, an audit trail and a scored test suite before it touches a customer.
- ✓ Human-in-the-loop approvals
- ✓ Eval suites & regression tests
- ✓ Rate, spend and scope limits
- ✓ Full action audit log
> 06
Operations
Someone has to own it at 2am. We monitor runs, alert on failure, retry safely and keep a changelog of what the agents did and why.
- ✓ Run monitoring & alerting
- ✓ Idempotent retries
- ✓ Cost and token tracking
- ✓ Monthly tuning reviews
> use.cases
Where It Pays For Itself.
Customer support agent
Answers from your own documentation, escalates to a human with full context, and writes the ticket back to the helpdesk. Deployed on WhatsApp, Telegram, Slack or web.
Sales & lead routing
Enrich an inbound lead, score it against your criteria, draft the first reply, create the CRM record and notify the right person in seconds instead of days.
Document intelligence
Invoices, contracts, statements and forms parsed into structured rows with confidence scores, and anything uncertain queued for a human to confirm.
Internal operations bot
A Slack or WhatsApp assistant that runs reports, checks stock, updates records and answers 'what is the status of X' without anyone opening four tabs.
Reporting & reconciliation
Scheduled pulls from payment, sales and ledger sources, reconciled automatically, with exceptions raised to a person rather than buried in a sheet.
Content & outreach pipelines
Research, drafting and scheduling with an approval gate, so volume goes up without your brand voice going sideways.
> tooling
Tool-Agnostic On Purpose.
We are not resellers. The platform is chosen after the workflow is understood - and we will tell you when the honest answer is a spreadsheet and a cron job.
> Conversational
WhatsApp Business API
Telegram Bot API
Slack apps- —Web & voice
> Orchestration
n8n
Make
Zapier- —Custom workers
> AI layer
- —Frontier LLM gateways
- —Agent frameworks
- —Vector search
- —Eval harnesses
> Data
Google Sheets
Airtable
PostgreSQL- —Warehouses
> Business systems
- —CRMs
- —Helpdesks
- —Payments & billing
- —ERP / inventory
> Delivery
Docker
Cloudflare
AWS
CI/CD- —Observability
> the.call
No-Code, Self-Hosted or Custom.
Use a no-code platform when
- — The workflow is linear and low-volume
- — Speed matters more than unit cost
- — The tools already have first-party connectors
- — You want to validate before investing
Self-host n8n when
- — Task volume makes per-run pricing painful
- — Data cannot leave your infrastructure
- — You need custom code steps mid-flow
- — You want to own the automation outright
Write real code when
- — Money, stock or identity is involved
- — Retries must be idempotent and provable
- — Latency budgets are tight
- — The logic is now a product, not a chore
> how.it.runs
Measured Before, Measured After.
01
Process mapping
We watch the work as it happens today and measure it - volume, handling time, error rate, cost per task.
02
Automation blueprint
A written design: what gets automated, what stays human, which tools, what it costs to run monthly.
03
Pilot on one flow
One high-volume workflow shipped end to end, with evals and an approval gate, so the value is measured not promised.
04
Scale and operate
Roll out the next flows, monitor every run, and tune monthly against the numbers we set at the start.
> questions
Straight Answers.
Will an AI agent make things up?
It can, which is why we do not let a model be the last word on anything that matters. Answers are grounded in your own content, actions are constrained to a defined tool set, anything money- or identity-related runs through deterministic code, and every agent ships with a scored test suite we re-run when prompts or models change.
Do we have to replace the tools we already use?
No. Most engagements connect what you already run - Sheets, Airtable, your CRM, your helpdesk, your database. We only propose replacing a tool when it is the thing actually blocking the workflow.
n8n, Make or Zapier - which one?
Zapier is fastest to stand up and the most expensive at volume. Make handles branching logic well at a middle price. Self-hosted n8n is the cheapest to run at scale and keeps data inside your own infrastructure, but somebody has to operate it. We recommend after seeing your run volume, not before.
How quickly does one see something working?
The first pilot workflow is typically live within two to four weeks of the blueprint, because we deliberately scope it to one process rather than a platform.
Who owns the automations?
You do. Workflows, prompts, code and infrastructure are handed over with documentation. If you stop working with us, nothing stops running.
Can this run on WhatsApp for our customers?
Yes - we have shipped a production WhatsApp commerce and support bot handling payments, wallet balances and human handoff. See the case study in our work.
> start.with.one.workflow
Automate The Expensive One First.
Tell us the process that eats the most hours. We will tell you what it costs to remove it.