Lead qualification
Asks your qualifying questions on the site, Messenger or Instagram, scores the answers, and books the call or hands the lead over.
AI & automation
AI agents for Montréal and Québec businesses
Agents that do the repetitive part of the job and ask before doing anything that matters.
An agent is an AI that can take actions — look something up, update a record, prepare a message — not just answer. That is useful and risky in equal measure, so we define exactly what each agent may do, what it must hand to a person, and how you can see everything it did.
01 · Trigger
A new lead writes in at 10 p.m. through the site chat
02 · Agent
It asks your qualifying questions, checks availability and fills in the CRM record
03 · Human check
Anything outside its rules — a discount, a complaint — goes to your team
04 · Result
Your team starts the morning with qualified leads and a summary of each
An example flow. Your steps, tools and approvals are mapped with you before anything is built.
In practice
What it looks like in a business like yours
Follow-ups and reminders
Prepares follow-up emails after a quote or a missed call, for your team to approve and send.
Back-office preparation
Gathers what a weekly report, a content calendar or a customer file needs, so your team only reviews and decides.
What's included
What the work covers
- A written job description for each agent: allowed actions, forbidden actions, tone and hand-off rules
- The tools the agent can use, each with its own limited permission (read the calendar, create a CRM record, draft an email)
- Human approval for messages to customers, prices, refunds and any other action you choose
- A log of every action and conversation, and a weekly review of errors during the first month
How we work
From the first call to results
Define the job
With the people who do the work today, we write what the agent does, what it never does and when it hands over.
Build and test
We build the agent on your data and test it against real past cases, including the awkward ones.
Supervised launch
It starts with an approval on every action; rules are loosened only where its record justifies it.
Review
We go through the logs with you, fix what went wrong and decide what it should take on next.
What you get
Deliverables
- A working agent in your channels and tools
- The agent's rules and hand-off policy in writing
- Action log and dashboard
- Monthly review of its results and errors
What we don't promise
An agent is not an employee and will not run your business on its own. It can be wrong, or be misled by an unusual message, which is why customer-facing actions stay behind an approval until you decide otherwise.
How it's priced
Quoted per mandate after a free call. What moves the price: the number of tools the agent uses, the channels it works in, the volume of conversations or tasks, and the model usage paid to the provider.
Every mandate is confirmed in a written quote (scope, deliverables, timeline, GST/QST) before any work starts.
Built in-house
We run this work on our own products first
Our own operation runs on agents: each company in the Krealo Publisher has its own AI agent that drafts and plans content, and nothing goes live until a person approves it. The agents we build for you follow the same rule.
- Active clients
- 25+
- Posts published in 90 days
- 4,000+
- Blog articles in 90 days
- 1,000+
Source: Krealo Publisher data for our client companies, July 14 – October 12, 2026.
FAQ
Questions, answered directly
When does an AI agent ask for human approval?
Whenever an action is on the list you set: messages to customers, prices or discounts, refunds, deletions, or anything unusual. At launch a person approves everything; rules are then relaxed one at a time, based on the agent's log.
What is the difference between an AI agent and a chatbot?
A chatbot answers questions in a conversation. An agent can also act: look up an order, create a CRM record, prepare an email or a report. If all you need is answers on your website, our AI chatbot service is the better fit.
What determines the budget of an AI agent project?
The number of tools and channels the agent works with, the volume of tasks, the level of supervision you want and the model usage costs. The written quote lists each of these before work starts.
What do we need to prepare before starting?
Examples of the work the agent will do (past emails, leads, tickets), access to the tools involved, and a person who decides what the agent may and may not do.
Related services
Often combined with
AI integration
AI that works inside your CRM, inbox and store — not in another tab your team forgets to open.
AI chatbots
A chatbot that answers from your catalog and your policies — and knows when to call a person.
Workflow automation
Stop copying the same data between five tools. Let the systems pass it to each other.