The brief for in-house counsel used to be clear: know the law, manage the risk, and keep the business out of trouble. But now, there is one more item on that list: figuring out AI, not as a future consideration, but as something a legal team is already using, or should be.
How Canadian in-house counsel are using AI day to day
The starting point for most in-house AI use is not strategy; it is a specific, painful task that takes too long.
For Melissa Reiter, general counsel and senior director at Jobber, an Alberta-based software company, that task was NDAs. Her team built an automation using existing internal tools, connecting spreadsheets to template documents, and cut NDA turnaround time from 24 hours to five to 10 minutes.
“A good piece of the automation has been using a tool that existed in the business and then building the pieces of it to kind of MacGyver it together,” she says. That DIY approach is more common than most vendor demos suggest.
Reiter’s team also uses AI to:
- generate first draft responses to internal business queries, drawing on a library of past legal advice
- auto-populate contract first drafts for routine matters
- track regulatory changes through legislative stages
The pattern across each use is the same: a repetitive, high-volume task gets a layer of AI, with a lawyer reviewing the output before anything goes out. The goal for in-house counsel using AI is not to replace the work. Rather, it is to “make space to focus on high value and strategic work.”
Building the governance layer: policy, tools, and oversight
Buying a tool and governing it are two different things. Most in-house legal departments in Canada are further along on the first.
Every in-house team needs to answer these questions before going further:
- Which tools are approved for legal use, and which are not?
- What data can go into an AI system?
- Who reviews AI output before it is relied on?
- How does the team stay current as tools and regulations change?
At Jobber, the team uses approved tooling within guardrails set by the business, with clear expectations about what can be built and what oversight is required. Legal operations holds this together; a lawyer on the team specifically focused on building workflows, testing tools, and maintaining the systems that keep everything running.
On the regulatory side, Canada’s picture is still forming:
- The Artificial Intelligence and Data Act (AIDA) did not proceed after Parliament was prorogued in January 2025
- Ontario’s Enhancing Digital Security and Trust Act (EDSTA) is in force but is still awaiting regulations
- Quebec’s Law 25 already requires organizations to disclose when decisions are made exclusively through automated processing
For a general counsel (GC) building an AI governance framework right now, that regulatory gap is both a risk and an opportunity, and there is room to set your own standard before one is set for you.
Check out this podcast from CL Talk for more insights on AI from the perspective of a senior legal counsel:
Head over to Lexpert’s In-House Lawyer page for more news dedicated for in-house lawyers and general counsel.
AI and outside counsel: who controls the relationship now
Something has shifted in how in-house legal departments relate to external law firms, and AI is certainly driving it.
In-house teams are using AI to reduce the volume of work that used to go outside. That changes what external counsel is actually for. At Jobber, the team has built internal depth in:
- fintech
- advertising and marketing
- commercial leasing
- labour and employment
- privacy
External counsel now comes in for highly specialized matters, regional expertise, or where someone needs to be, as Reiter puts it, “connecting with regulators.”
“Especially in specialized areas, we work with some people who are living and breathing this, connecting with regulators — I think in the span of 10 minutes — can say, oh yeah, I’m right on top of that issue,” she says.
AI is not replacing external counsel; it is raising the bar for what justifies sending work out. Routine matters that a well-governed AI workflow can handle in minutes are harder to justify billing at hourly rates.
The GCs who are getting real value from outside firms are the ones being specific about what they actually need: deep expertise, regulatory access, and accountability — not capacity.
What GCs need to get right before scaling AI further
Scaling AI in-house counsel operations is not a technology problem. It is a people and process problem with a technology layer on top. Three things matter most:
1. Hire for curiosity, not just credentials
Reiter wanted a lawyer who saw legal operations as an opportunity, not an obligation. “I absolutely would look at fluency with AI and a level of comfort using that as part of the things that you do each day,” she says. As such, openness to learning counts as much as current skill.
2. Question your own processes before adding AI to them
“Can we be doing things faster or better? Can we scrap processes that no longer serve us?” Reiter asks. Scaling AI on top of a broken process produces faster broken outcomes. The process question must come before the tool.
3. Stay deliberate when things go sideways
Things will go wrong. For instance, an AI tool may be used in ways it was not built for, or an output may be cleared when it should not have been. Reiter’s approach to high-stakes situations applies here too: “slow is smooth and smooth is fast.” A careful, planned response beats moving quickly in the wrong direction.
The decisions about what gets automated, what oversight looks like, and where the team is headed all belong to the person running the department.
Subscribe to the free Lexpert newsletter for more articles to help general counsel and in-house lawyers, including topics on AI tools for in-house counsel.