Seven years ago, two Laivly developers took the stage at Winnipeg’s Prairie Dev Con to make the case for a bold idea: that machine learning could understand customer conversations like a person would. And this year, Laivly was back at Prairie Dev Con to open the show.
Prairie Dev Con made its annual return to Winnipeg September 21–22, 2026, bringing together developers looking to sharpen their skills and figure out how new technology fits into their work. Craig Shyjak, Laivly’s VP of Architecture, delivered the keynote, “Generation Is Cheap. Judgment Isn’t.”
Craig has spent more than 20 years building and leading teams behind scalable platforms across cloud, data, and AI. Lately, his focus has shifted to agentic AI and what it means for how software gets built, from AI-driven developer workflows to systems where agents act on a user’s behalf. His talk looks at what happens as AI takes on more of the work of writing code, reviewing pull requests, and even diagnosing production failures, and digs into where human judgment still matters and who owns the risk when the AI gets it wrong. Neither anti-AI nor AI hype, Craig argues for using these tools with intention and grounds it in what it actually takes to turn emerging capability into production-ready systems.
The Early Days of Sidd
Back in 2019, lead software developers Josh McPhail and Jordan Cole—now Director of Sales Engineering and Principal Machine Learning Engineer, respectively—presented at the same conference. Their focus was on actionable detection and natural language processing, which was the early thinking behind Sidd, Laivly’s agentic AI and agent assist tool.
Jordan, the first developer at Laivly to pioneer the Sidd project, noted at the time that machine learning had already been around for decades but had finally reached a point where it was ready for commercial use. And when asked about disruptive technologies, Josh explained that it was all a matter of optimization: use the technology to free up people to take on the next challenge.
That thinking hasn’t changed, but it has gotten more urgent. Today, Sidd is a full ecosystem that handles customer cases from end to end, creates smooth handoffs from AI to human with full context, and helps agents provide faster, more accurate, and more attentive customer experiences, all thanks to the very capabilities Josh and Jordan presented on seven years ago.
A Laivly Presence on the Prairies
The 2026 Prairie Dev Con drew more than 40 speakers and 70 sessions to Winnipeg, and Laivly was well represented.
Craig opened the conference, giving his keynote to a packed house of engineering professionals. He touched on everything from vibe coding to fully agentic engineering, and made the case that as AI makes code cheaper to produce, human judgment and ownership matter more than ever.
Later, Addison Rennick, Principal Software Developer at Laivly, took the stage with a session on “Speed vs. Accuracy: Execution Modes for Agentic Workflows,” breaking down how to balance fast execution with careful execution when putting AI agents to work.
Seven years ago, Josh and Jordan presented on actionable detection and NLP, and this year, Craig and Addison picked up that conversation and carried it forward.
