AI & Automation Services
Useful, or not at all.
AI and automation services cover the work of deciding where software can take a repetitive task off a person, then building and running it: assessment, workflow automation, custom development, AI agents, integration with the systems you already own, Microsoft 365 Copilot, and the policy underneath. happier IT does that work for Canadian organizations of roughly 15 to 200 people in Alberta, British Columbia and Ontario.
Who it's for
Most of what people call an AI problem is not one.
Four conversations bring people here. Two of them end in a build, and two end in us saying the money would do more good somewhere else first.
There is a job everyone hates and nobody owns. Retyping figures from PDF invoices, chasing the same approval every Thursday, copying data between two systems that will not speak to each other. That is usually a genuine automation win, and it often needs no AI at all.
Someone senior has asked about Copilot. A licence quote arrived, the demonstration was good, and nobody is sure whether it would work here. That question has a real answer and it is mostly about your file permissions. See Microsoft 365 Copilot deployment.
Staff are already using AI tools nobody approved. Free accounts, personal logins, work documents pasted in. This is not a discipline problem. The tools are useful and nobody wrote down which ones were allowed.
The data is the problem, not the tooling. Three spellings of the same customer name, a spreadsheet that is really a database, a process that changes depending on who is doing it. No model fixes that. It has to be fixed first, and it is ordinary work.
What we will not tell you
We are not going to say AI will transform your business, or quote you a productivity figure. We do not have a sourced one for an organization your size, and neither does anyone showing you a slide with a percentage on it.
What we will do is name the specific tasks worth automating in your specific business, and say out loud which ones are not.
What's included
The seven things we actually do.
Each of these has its own page, because each is bought separately and answered differently. Start wherever your question is.
-
AI readiness assessment
A short structured look at your data, permissions, processes and licensing, ending in a written answer about what is worth doing and in what order. <a href="/ai/readiness-assessment/">AI readiness assessment</a>.
-
Business process automation
Workflow automation: software that carries out a sequence of steps on its own, on a trigger, without a person moving the work along. Approvals, form-to-system handoffs, scheduled reports. <a href="/ai/automation/">Business process automation</a>.
-
Custom AI development
Building something that does not exist off the shelf: a tool that answers questions from your own documents, a classifier for incoming work, a model tuned to your terminology. <a href="/ai/custom-development/">Custom AI development</a>.
-
AI agents and assistants
An agent is software given a goal, a set of tools and permission to take steps on its own until it is done, as opposed to a chat window that only answers. Narrow ones work. <a href="/ai/ai-agents/">AI agents and assistants</a>.
-
AI integration services
Connecting AI to the systems that hold your real data, your accounting package, CRM, ERP, ticketing or field-service software, through their APIs, so nobody has to copy and paste. <a href="/ai/integrations/">AI integration services</a>.
-
Microsoft 365 Copilot deployment
Readiness, permissions clean-up, licensing advice, a measured pilot, and the training that decides whether anybody keeps using it. <a href="/ai/microsoft-copilot/">Microsoft 365 Copilot deployment</a>.
-
AI governance and policy
A one-page policy people will actually read, an approved-tool list, access managed like any other system, and honest answers on data residency, where your information is processed and stored. <a href="/ai/governance/">AI governance and policy</a>.
-
And the ordinary IT underneath all of it
Identity, permissions, backup and monitoring do not stop mattering because the project is called AI. They matter more, because an automation runs unattended. That work sits in <a href="/managed-it-services/">managed IT</a>.
How it works
How an engagement usually runs.
Smallest useful thing first. A pilot that works beats a strategy document, and it costs less to be wrong.
-
Find the tasks, not the technology
We sit with the people doing the work and write down what actually takes time: the steps, the systems, the exceptions, and how often the exception happens. Volume and consistency decide what is worth automating. Enthusiasm does not.
-
Build one thing and measure it
One workflow, one team, an agreed measure of whether it helped, and a date to decide. Small enough that being wrong costs weeks rather than a budget year, and specific enough that the answer is not a matter of opinion.
-
Run it, own it, or hand it over
Anything we build gets a named owner, monitoring, and documentation good enough for somebody else to maintain. If you would rather run it in-house, we hand it over properly rather than keeping you tied to us.
What it costs
Assessment fixed, build quoted, running included.
Three different shapes of cost, because they are three different kinds of work. Anyone quoting all of it as one monthly number is guessing.
The assessment It is deliberately something you can buy without committing to a build.
Automation and development
Licences come from the vendor and are usually per user, per month. Microsoft 365 Copilot Chat is included at no additional cost with eligible Microsoft 365 plans; the Microsoft 365 Copilot that reads your own files and email is a paid add-on.
Running what has been built, monitoring, updates, the call when something breaks at 7am, sits inside the managed IT
The cost nobody quotes
Every automation has a running cost in attention: someone has to notice when it stops, know what it was for, and be able to change it when the process changes.
A flow built by one enthusiastic person who then moves teams is the most common failure we are asked to clean up. Ask any provider who owns each thing they build, and what happens when that person leaves.
Go deeper
Specialisms under this service
Each of these is a distinct piece of work with its own page. Most clients need some of them, not all of them.
AI Readiness Assessment
Before you buy anything.
A structured review of your data, permissions, processes and licensing, ending in a written answer on what AI is worth doing here, and what is not yet.
Business Process Automation
The boring wins first.
Workflow automation for Canadian organizations of 15 to 200 people: approvals, form-to-system handoffs, scheduled reports, and the copy-paste jobs.
Custom AI Development
Built, when bought will not do.
Custom AI builds for Canadian organizations: retrieval over your own documents, classification, document extraction, and honest advice on when to buy instead.
AI Agents & Assistants
Narrow jobs, done the same way.
AI agents built narrow, permissioned and monitored, for Canadian organizations of 15 to 200 people. What they do well, and where a workflow is the better answer.
AI Integration Services
Where your data already lives.
Connecting AI to the systems you already run, accounting, CRM, ERP, ticketing, field service, through their APIs, so nobody copies and pastes between them.
Microsoft 365 Copilot Deployment
Permissions first, licences second.
Microsoft 365 Copilot for Canadian teams of 15 to 200: readiness, permissions clean-up, licensing advice, a measured pilot, and the training that lands.
AI Governance & Policy
One page, actually read.
An AI policy people will read, an approved-tool list, access managed properly, and straight answers on Canadian privacy law and where your data is processed.
Why us for this
What we can honestly say, and what we cannot.
happier IT has been building and running IT for Canadian organizations for years, and automation work is a large part of that. Three of the things we have built: connections into software APIs that pull real-time information from several locations into one view, candidate processing for a human resources team, and automated skills coaching for professional development.
What we can say plainly: we run our own security operations centre in Canada, staffed by our own employees, which is where unsanctioned tools and unusual data movement become visible. Staffed hours: 24/7, with the Surrey office on Monday to Friday, 8:00 am to 5:00 pm Pacific. Our Microsoft certifications are listed in full on our awards and certifications page.
And the honest limit of the whole category: at 15 to 200 people, AI does not replace roles. It removes specific repeated tasks from specific people’s days. That is worth real money and it is a much smaller claim than the one being made at you elsewhere.
Go deeper
- AI readiness assessment The nine-question version takes about a minute.
- AI governance for Canadian SMBs The seven decisions to make before staff paste anything in.
- Safe automation for manufacturers A worked example, done carefully.
Questions
What people ask before they sign anything.
What can AI actually do for a company of 40 people?
Four things reliably: drafting first versions of routine writing, summarising long documents and meetings, searching your own files in plain language, and taking a repetitive copy-between-systems job off somebody entirely. The last one is usually plain workflow automation rather than AI, and it is often the biggest single win. What it will not do at your size is replace a role, or fix data that was inconsistent before you started.
When is AI the wrong answer?
When the underlying problem is a process problem or a data problem. If the same task is done three different ways depending on who picks it up, automating it just makes the inconsistency faster. If your customer records have three spellings of the same company, a model will confidently answer from the wrong one. Both are fixable, both are ordinary work, and both have to come first. We will say so before quoting anything.
Do we need Microsoft 365 Copilot or something custom?
Start with Copilot if your work happens in Microsoft 365 and your file permissions are tidy, because it needs no build. Consider something custom when the useful knowledge lives outside Microsoft 365, in a line-of-business system, a document archive, a database, or when you need a specific repeated task done the same way every time. Most organizations of this size need Copilot plus two or three small automations, not a platform.
Is our data safe if we use AI tools?
It depends entirely on the tool and the tier, which is why the answer has to be per product rather than in general. Ask every vendor four questions: is our data used to train their models, where is it processed, where is it stored at rest, and how long is it kept. Microsoft states that Microsoft 365 Copilot prompts, responses and data retrieved through Microsoft Graph are not used to train its foundation models. Consumer accounts on other services often say something different. AI governance goes through this properly.
What does AI and automation work cost?
Three separate numbers. Builds are quoted per project against a written scope. Licences are per user, per month from the vendor. Beware of any quote that folds all three into one monthly figure, it hides which part you could stop paying for.
Can you build something custom, or do you only resell Microsoft?
We build. Custom development, integrations between systems that were never designed to talk to each other, and automations that run unattended are a normal part of the work, see custom AI development. Where an off-the-shelf product already does the job, we will tell you to buy it instead, because a bought thing has a support line and a built thing has us.
Is there a law in Canada about business use of AI?
There is no comprehensive federal AI statute in force. The Artificial Intelligence and Data Act, part of Bill C-27, died when Parliament was prorogued on 5 January 2025. What does apply is everything that already applied: federal and provincial privacy law, your contracts, sector regulators, and human rights and consumer protection law. Canada’s privacy commissioners have also published joint principles for generative AI. Details on AI governance and policy.
Related
Where to go next.
Want to know what this would look like for you?
A 30-minute call. No slides, no audit fee, no obligation. We ask what is breaking and tell you honestly whether we are the right fit.