Reflections from PEI on regulatory pressure, merger complexity and making modernization useful for credit unions and their members.
We went to the Canadian Credit Union Association 2026 Financial Forum in Charlottetown, PEI, to listen and learn. Our team wanted to hear firsthand how credit unions are navigating their technology priorities and what realistic, incremental progress might look like.
Across our conversations, one theme emerged: credit unions are managing significant change, yet much of the work to accommodate it remains manual.
Regulatory requirements evolve. Organizations merge. Products and processes must align. AI introduces new possibilities but also another set of decisions about where to spend time and money.
Each development brings its own demands. Together, they place pressure on the people responsible for keeping services running and members supported.
Seeing opportunity in the face of threat
One standout moment was the keynote by Graham Sherman, founder of Tool Shed Brewing Company. His message about innovation, resilience and adaptability was especially well received. Our takeaway:
Look for opportunity in the face of threat, challenge familiar assumptions and meet disruption with creativity, grit and determination.
As CCUA highlighted in its recap, Sherman challenged the tendency to default to “this is the way it’s always been done.” His entrepreneurial journey reminded us that questioning established approaches can open up possibilities for an entire industry.
For us, that message connected directly to the conversations throughout the Forum. Credit unions support entrepreneurs who continually adapt to uncertainty, and their own organizations face decisions about what to preserve, what to rethink and where to try something different.
Behind a lending application is a business trying to move forward. Behind a delayed approval may be an opportunity with a deadline. The processes supporting those relationships influence how effectively a credit union can respond.
Sherman’s challenge applies to those of us supporting the sector, too. A duplicated workflow, a difficult handoff or a process inherited through a merger deserves a fresh look. Asking why it works that way, and giving people room to imagine a better approach, can be the first practical step toward change.
That invitation to question familiar approaches came to mind during several of our conversations about AI.
AI is on the desktop. Connecting it to the business is the next step.
Nearly everyone we spoke with had Microsoft Copilot on their workstation and used it for productivity tasks such as summarizing documents. But when we asked whether it would help to connect Copilot to reporting, document management and line-of-business systems, without manually downloading and uploading files, the response was revealing:
“Is it possible to do that?”
That exchange highlighted a gap between everyday AI use and awareness of its potential when connected to enterprise systems. This reflects our conversations, rather than a sector-wide assessment.
The opportunity extends beyond summarizing documents: bringing together policies, organizational data and the context of a member’s request or an employee’s task to support better-informed decisions and actions.
Realizing that opportunity requires suitable data, reliable integrations and appropriate permissions. Budget and staff capacity also matter. A focused pilot, built around one task with measurable value, can help establish the effort and controls required before expanding.
Forecasting makes those connections tangible
Forecasting attracted particular interest in our conversations. Liquidity forecasting provides a useful example of why the surrounding systems matter.
Depending on the scope, a forecast may need information about deposit behaviour, expected loan drawdowns, scheduled payments and funding commitments. Those inputs may sit across core banking, lending, payments and treasury systems, each with different definitions and update cycles.
The design work starts with understanding which sources are needed, how they connect, and whether the information is sufficiently complete and timely to support the decision.
Then evaluate the forecasting method against the current approach. What would a better forecast help someone do differently? How will accuracy and uncertainty be assessed? What happens when the forecast is wrong?
Those connections also need clear controls. Teams should be able to identify the data, assumptions, and model version behind a forecast; record adjustments and approvals; and understand what action followed. Timely monitoring can flag stale inputs or results that need review.
These are considerations for a potential application, rather than a claim of proven results. We believe the value comes from connecting reliable analysis to a decision someone can responsibly make.
AI agents need clear workflows and controls
Connecting AI to information and authorizing it to act are separate decisions. An agent might interpret a member’s request, gather supporting information or recommend a next step. The surrounding workflow should define the applicable rules, mandatory checks, required approvals, and decision ownership.
Agents can contribute intelligence to specific tasks without redefining the process each time.
For credit unions, meaningful adoption therefore requires thinking about both system connections and how work gets done with clear ownership, human review where needed and a traceable record of actions.
There is still a lot of manual work
The amount of manual processing we discussed was striking.
Manual work is not inherently a problem. Member service and sound decision-making depend on human judgment. But there is a difference between exercising judgment and spending time moving information, chasing updates or coordinating the next step.
Consider a request that begins in one system but needs an email to reach the next team. Or an approval whose status is tracked separately from the application it relates to. People involved may make that process work reliably through experience and effort, but that effort consumes capacity and affects the member experience.
For us, a useful starting question is: where does work leave the system, and what must an employee do to keep it moving? Following that path often reveals practical opportunities to connect information, clarify ownership and reduce repeated handling.
A merger continues long after the organizations come together
Merger activity was another prominent theme in our discussions, alongside expectations of further consolidation. CCUA’s discussion of consolidation reflects the broader context.
The complexity extends into everyday service. Teams within the same credit union may follow different processes for the same product. Workflows may overlap. Employees may need to know which approach applies, while members experience different outcomes depending on where or how they access a service.
That creates frustration on both sides of the counter.
Our takeaway is that process integration deserves as much attention as systems integration. Bringing organizations together requires understanding what exists, deciding what should be consistent, and identifying where differences remain justified.
Automating too early can preserve the duplication. A better starting point is to map the existing approaches with the people who use them, agree on the intended process and then establish how systems and automation should support it.
The practical test is whether employees can work confidently across the combined organization and whether members experience a coherent service.
Regulatory change becomes operational work
One concern we heard was the challenge of applying frequent regulatory changes and the risk of getting implementation wrong.
FINTRAC’s updates affecting reporting entities illustrate the implementation work institutions need to track. A new requirement must travel a long way before it becomes standard practice. Someone must interpret it, identify the affected products and processes, update procedures, communicate the changes and establish how compliance will be demonstrated.
That becomes harder when the same activity happens differently across teams or locations.
From our perspective, this is where regulatory change becomes a workflow challenge. An updated policy needs a clear path into the forms people complete, the approvals they request, the checks they perform and the evidence they retain.
A related reflection from our team was that risk and compliance can be treated as assumed foundations, or ‘table stakes,’ without enough attention to how they operate within a particular process. They deserve explicit consideration in the design of both the workflow and the technology supporting it.
That means being able to reconstruct a past decision: what information was available, which rules applied, who approved it and whether an override occurred. It also means having timely visibility into work happening now: outstanding reviews, exceptions and missed control steps.
Historical auditability and current oversight serve different purposes. Both need to be designed into the process, with appropriate records, access controls and monitoring.
Technology can support that implementation. It still requires people with the expertise to determine what a requirement means and how it should apply.
Process gaps can extend beyond the credit union
An extended conversation with one credit union raised an opportunity worth investigating: how information about a home insurance claim reaches the credit union holding the mortgage.
The concern was a potential disconnect between insurer, homeowner and lender. Missing or delayed notifications could affect coordination, including how claim proceeds are handled in relation to repairs and the outstanding mortgage.
The circumstances, notification obligations and payment arrangements would need validation. One conversation does not establish a sector-wide problem.
But it raises a useful question: could a shared portal or integration service let participating insurers submit relevant notifications securely and route them to the right credit union?
Each notification could trigger a defined process with ownership, acknowledgement, follow-up, and an audit trail. Discovery would need to identify where existing arrangements break down and what information can be shared appropriately.
It is a concrete example of the potential to improve coordination across organizational boundaries.
Continuing the conversation
The questions we brought home from PEI are practical ones. How do you apply a new requirement consistently? How do you bring different processes together after a merger? Give employees more time for members? Improve coordination with partners? Connect the right systems for a useful AI application, with controls that work throughout the process?
These are the kinds of challenges we at AOT Technologies help organizations work through. Progress often begins with one process: understanding where it becomes difficult, agreeing on what better looks like and finding a manageable place to start.
Graham Sherman’s keynote encouraged us to see opportunity in the challenges ahead. If these observations reflect what is happening in your credit union, we’d welcome a chance to compare notes on what you’re working through, what you’d like to make possible and where a useful first step might be.
Author Profile:
Moray Hickes is an Enterprise Technology Solutions leader with over two decades of experience helping organizations navigate complex technology systems and deliver high-value solutions across regulated industries.