Automation and AI
Automation and AI, grounded in your real data.
Metiva automates the routine work (follow-up sequences, routing, status changes) with a simple trigger, condition, action builder, and helps your team reply faster with AI assist inside the inbox. Because it all runs on one clean identity spine, the AI works from your real records, not a bolt-on guess.
30 minutes, no slides. We map your current stack live. Dutch or English.
Hi Jeroen, donderdag kan zeker. Schikt 14:00, of zal ik een paar opties sturen?
A person reads and approves every send
An AI-assisted reply drafted in the inbox, with sample data, illustrative.
Why this AI is different
AI is only as good as the data underneath it.
Most agency AI is bolted onto a stack of disconnected tools, so it guesses from fragments: a chat in one app, a CRM record in another, an ad platform that never reconciles. Metiva is one system. Every message, lead, deal, and event resolves to one person record on one spine. So when an automation fires or the AI drafts a reply, it is working from the whole, current truth, not a copy that drifted overnight.
- One identity spine. Every touch resolves to one canonical person, so the AI sees one history, not three partial ones.
- Clean, first-party events. The same events that power server-side attribution also tell automation exactly when to act.
- Shared by design. The inbox, the pipeline, and the reporting are the same data, so there is nothing to sync and nothing to drift.
Automation builder
Build a rule the way you would say it out loud.
Routine steps stop being manual. A trigger, condition, action builder lets you describe what should happen in plain terms: when this arrives, if it matches, do that. No code, no separate automation tool, and it acts on the same records the inbox, pipeline, and reporting already use.
- Triggers on real events: a new lead, an inbound message on any channel, a stage change, a signed contract, or an overdue invoice.
- Conditions on real fields: channel, source, deal value, lifecycle stage, owner, or score, so a rule only fires where you mean it to.
- Actions that move work: assign an owner, set a stage, add a note, send a templated reply, or start a follow-up sequence. Each rule runs once per record so it never loops.
This is rule-driven and template-driven automation, kept simple on purpose. You can require a person to approve any outbound step, so the system never sends behind your back.
- Welkom + korte introEmailDay 0
- Korte check-inWhatsAppDay 2
- Laatste reminderEmailDay 5
Follow-up sequences
Follow-ups that send themselves, on your timing.
A lead goes quiet, a proposal sits unsigned, a client has not heard from you in a month. A follow-up sequence handles the reminder so a person does not have to remember. Sequences are built from your own templates and run on the same records, so they stop the moment a lead replies or a deal moves.
- Multi-step email and message sequences from reusable templates with merge tags, timed in hours or days.
- Enrolment by rule (a new lead, a stalled deal, an unsigned contract) and automatic exit the moment the contact responds or the stage changes.
- Every send is logged against the one person record, so the sequence and the live conversation never contradict each other.
This is template-driven and rule-driven follow-up, not a heavy campaign-nurture suite. We keep it simple on purpose, and consent state is respected on every send.
AI assist
Draft, rewrite, translate, and summarize, without leaving the thread.
AI assist lives inside the inbox composer to speed up the reply your agent is already writing. It works from the real conversation and the linked record, so the suggestion fits the client in front of you. Your team reads and approves every send.
- Seven actions in the composer: draft, rewrite, expand, shorten, translate, and summarize a reply, and suggest the next pipeline stage.
- Grounded in the open thread and the linked deal and history, so a draft is about this client, not a generic template.
- A person approves every message before it sends. In the inbox, the AI never answers a client on its own.
AI assist is gated behind your configuration and an API key, and it never sends without a person.
Hi Jeroen, donderdag kan zeker. Schikt 14:00, of zal ik een paar opties sturen?
A person reads and approves every send
The AI chatbot
The chatbot ships last. That is the point.
A first-line chatbot is only trustworthy if it answers from the truth. Bolt-on bots are wired to a stack that does not agree with itself, so they guess, and they invent. Metiva's chatbot is built to ship after the identity spine and the clean events are in place. When it does answer a website visitor, it reads the same records your team reads. It captures the lead onto one record, answers from your real knowledge, and hands off to a person the moment it should.
- Grounded in your own records and knowledge, so it answers from what is true for your agency, not a generic model guess.
- Captures the visitor onto one canonical person record from the first message, so the conversation, the lead, and any later deal are one history.
- Escalates to a teammate with full context when it is unsure or when the visitor asks, and never fabricates an answer to fill a gap.
The AI chatbot is on the rollout roadmap and we configure it with you once your data foundation is in place. We would rather ship it late and grounded than early and wrong. It is a separate capability from the AI assist your team already uses in the inbox.
Grounded by design
Grounded, supervised, and consent-aware.
AI is useful here because it is fenced in. It answers from your data, a person stays in control of what reaches a client, and the same consent and EU-data rules that govern the rest of Metiva govern the AI too.
Grounded in your data
The AI works from your real records and knowledge, not a generic guess. If it does not know, it says so or escalates instead of inventing.
A person stays in control
In the inbox, every AI draft is approved by an agent before it sends. Automations can require human approval on any outbound step.
Consent and AVG respected
Sequences and messages honor consent state, and personal data is handled under the same deny-by-default rules as the rest of the system.
Your data stays in the EU
AI features run on your tenant's data with EU residency, and each client's data is isolated by design and fail-closed.
No autonomous sends in the inbox, no answers invented to fill a gap. The AI runs on your tenant's data, under the same consent and isolation rules as the rest of Metiva.
Grounded, not guessed
Counts, not promises.
These are capability counts and design choices from the product, not performance promises. Your numbers are what we map together on the call.
Questions
Straight answers about automation and AI.
Metiva is one system, priced per company, not per seat: a one-time implementation, then a monthly service retainer priced by the client brands you manage, with unlimited seats. The monthly is a service retainer, not software rental: hosting, maintained integrations, support, and accurate attribution. Automation and AI assist are part of the system, not a separate add-on fee. 12-month commitment, then cancel monthly, with two months free on annual prepay. Tell us how you run and we quote you directly, on WhatsApp or by email.
In the inbox, no. AI assist works inside the composer to help your team draft, rewrite, translate, and summarize a reply, and to suggest the next stage. A person approves every send. The autonomous first-line website chatbot is a separate capability on the rollout roadmap, and even then it answers only from your real data and hands off to a human when it should.
From your own system. Because every message, lead, and deal resolves to one person record on one spine, the AI works from your real records and knowledge, not a generic model guess. That is the whole reason it is trustworthy here.
Yes. You describe a rule as a trigger, a condition, and an action: when this happens, if it matches, do that. No code and no separate automation tool. Rules act on the same records the rest of Metiva uses, and you can require a person to approve any outbound step.
It is on the rollout roadmap, by design. The chatbot is built to ship after your identity spine and clean events are in place, so it can answer from the truth instead of guessing. We configure it with you when your data foundation is ready. The AI assist your team uses in the inbox is available now.
Yes. AI features run on your tenant's data with EU residency. Consent is deny-by-default, personal data is handled under the same rules as the rest of the system, and each client's data is isolated by design and fail-closed. A person stays in control of what the AI sends.
One connected system
Put the routine on autopilot, keep your judgment in the loop.
The automation, the AI assist, and the chatbot all run on one clean system, so they work from the same truth your team does. We will show you the builder and the composer with your real stack on the screen.
30 minutes. No slides. We map your current stack live. Dutch or English.