The Canadian tech and professional lines market is loosening after several hard years, but underwriters are carving out firmer pockets for managed service providers (MSPs), SaaS firms and true AI exposures, according to SWG's Peter Kyung Min Kim (pictured).
Kim, senior professional lines underwriter at SWG, said the segment has shifted noticeably from the conditions he saw just a few years ago.
“It’s softening,” he said. “It’s moving from a hard market compared to a couple of years ago. From what I’m seeing, the market has been very aggressive in trying to grow more, so capacity is there.”
He pointed to the arrival of new, smaller MGAs as adding further competitive pressure, particularly in core technology and professional classes. That has translated into more quotes on the table for brokers and clients and, in many cases, more flexibility on terms.
With the rise of new, smaller MGAs on the horizon, that creates competition among MGAs with the brokers, he said.
“That’s good news for insured consumers – there’s more quotation out there so you can compare with other markets," he said.
Kim stressed, however, that the softening is not uniform. Accounts with weak controls, adverse loss histories or unusual exposures are still likely to face firm pricing and tighter wordings.
“From my experience, firms with weak control, loss histories, or if your operation involves niche services – let’s say involving AI – those types of risks still face firm pricing and tighter terms,” he said.
MSPs and SaaS platforms sit close to that line. Both can be business-critical for clients and can create complex chains of dependency if something goes wrong. Kim said he is seeing more of these submissions flagged for deeper scrutiny.
The term “AI” itself is now one of the most common triggers for closer review. Kim said it is also one of the most commonly misused labels in submissions.
“The term AI is being used popularly these days,” he said. “But what we’re really concerned about is the use of AI.”
He drew a sharp distinction between tools that use algorithms to enhance productivity – which have existed in various forms for years – and systems that are genuinely autonomous.
“What we’re more concerned about is the AI that usually has an autonomous system that has learning capability, that has its own decision‑making abilities,” he said. “Those are the type of risk that we are looking at more closely.”
As an example, he contrasted straightforward automation in an office environment with autonomous control in a production setting. A brokerage that uses software to triage renewals based on premium, class or revenue is a different risk from a manufacturer whose plant equipment is being optimised in real time by a learning system. A failure in the latter can lead directly to stoppages and significant financial loss.
Kim said underwriters are responding by building more granular intake frameworks. Standard tech E&O applications are no longer sufficient to capture the full picture for MSP, SaaS or AI‑driven operations.
“We’re now seeing dedicated questionnaire for Managed Service Providers, AI, SaaS driven operations,” he said. “Your standard application can no longer capture the full detail of those operations.”
While he recognises the burden this places on brokers, particularly those juggling multiple markets and forms, he argued that the extra information is necessary.
“Markets are also adjusting to the risks they see – we want a deeper understanding of risks we see,” Kim said. “Sometimes we don’t get the full picture right away, so there’s going back and forth between the broker and the underwriter, trying to figure out where this risk really lies and if this is truly an MSP or AI exposure.”
One key red flag, he said, is the presence of fully autonomous decision‑making with no human oversight or governance. Questionnaires are probing whether there is any formal validation process, and whether tools are for internal use only, advisory use to support humans, or constitute final decision‑makers.
“If the AI is truly a fully autonomous decision maker with no human oversight, that is a red flag from an underwriting perspective,” he said. By contrast, advisory or internal productivity tools, used under human control, are generally viewed more favourably.
The result is a two‑track market. Conventional IT services, consultants and lower‑hazard tech firms are benefiting from greater capacity and easing terms, while firms with complex dependency chains or genuinely autonomous systems find themselves under a brighter spotlight.
