Issue #2 · March 2026
March brought some of the clearest signals yet that AI isn't just changing which tools we use, it's changing how work itself gets structured. Meanwhile, the economics of building software shifted in ways that would have been hard to believe even six months ago. This issue digs into what's changing on the ground: from the invisible costs hiding inside your operations, to the new math of building vs. buying software, to a competitive landscape that's moving fast.
Every organization has processes that quietly consume far more time, money, and human energy than they should. Data gets entered twice into different systems. Critical steps depend on one person who just happens to know how things work. People build workarounds because the tools don't match how the work actually happens. None of it is dramatic on its own, but the cost compounds. Extra hours, slower decisions, good people spending their days on admin instead of the work they were hired to do.
The challenge is that these costs are invisible until someone maps the full picture. That's what we've been doing in service blueprinting workshops with clients over the past few months: putting teams in a room and mapping a core process end-to-end from two perspectives, the people doing the work and the customer being served. We trace every action, system, handoff, and data flow. Without both lenses, it's easy to focus only on internal operations and miss the job the process exists to do for the people it serves.
“I've worked here for eight years and I've never seen the full picture of how this process actually works. No wonder we've been frustrated.”
Want to learn more about this organization and how their story unfolded? Reach us at hello@senseandmotion.co and we'll share the details.
What makes this especially powerful right now is that AI has changed what you can do once you know where a process breaks down. At a BC non-profit, our blueprinting revealed an estimated $150,000+ per year in administrative overhead from a broken system, evidence that enabled a clear business case for an AI-enabled replacement. At a fleet management company that we work with, three days of intensive sessions produced the process definitions that became the foundation for an entirely new department.
Most organizations don't need more AI tools. They need to understand their own processes well enough to know where AI would actually make a difference. If you're not sure where to start, a Process Discovery Workshop is a good way to find out.
To schedule a Process Discovery Workshop, email us at hello@senseandmotion.co.
For most organizations, building custom software has always been expensive and slow enough that buying off-the-shelf tools was the sensible choice, even when those tools didn't quite fit. That calculus is shifting fast. AI-assisted development has compressed the cost and timeline of building software so dramatically that examples from this month alone would have seemed implausible a year ago.
Cloudflare had a single engineer rebuild the most widely used React framework (Next.js, used by millions of developers) from scratch in one week using an AI coding agent, for $1,100 in AI tokens. It's still experimental, but something that previously would have required a large team and years of engineering effort was done in days. In a similar vein, OpenAI is reportedly building its own version of GitHub (the platform where most of the world's software code is stored and managed) after frustrations with recent service disruptions. When companies can stand up major software infrastructure this quickly, the old assumptions about what's worth building don't hold anymore.
For BC businesses, the "should we build or buy?" question looks completely different than it did six months ago. Off-the-shelf tools that don't quite fit your workflow may no longer be the pragmatic default.
How Cloudflare rebuilt Next.js with AI in one week →
OpenAI, Google, and Microsoft tend to dominate the AI conversation. But while everyone's been watching those three, Anthropic has been quietly building a commanding position. The company behind Claude is surging forward in share of enterprise AI spend, especially in API usage, and is becoming the preferred choice among businesses purchasing AI for the first time. At the same time, OpenAI is losing ground, with an enterprise spending share decline of nearly 50% over the last two years.
The enterprise AI market is consolidating around Claude faster than most people realize. At Sense & Motion, we've used Claude as our primary AI tool for over two years now and have watched it grow into the market leader in business productivity for much of that time. We also love Anthropic's safety-first mission. If your organization is evaluating AI providers or locking in partnerships, this is worth watching closely. The company winning the business market isn't the one with the most brand recognition.
We're a member of Anthropic's Claude Partner Network and have deep expertise across Claude's product suite, from Cowork to Code to the API. If you're looking to get your team up to speed, our AI Literacy for Claude program is a good place to start.
If the Cloudflare story above signals what's possible, what's happening inside major companies signals what's already real. AI is no longer just assisting developers, it's starting to write the majority of code itself. Uber's CTO shared that 95% of their engineers now use AI tools every month, while some organizations are writing 100% of their code with AI.
The speed is real, but so are the costs. "Token spend" (the cost of AI usage) is becoming a new line item that companies are learning to track. Engineers running AI agents can burn through millions of tokens in a day. Companies are now measuring AI consumption the way they used to measure cloud computing costs. Meta is now factoring AI usage into performance reviews, making "AI-driven impact" a core expectation for every employee starting this year.
And then there are the risks that don't show up on a bill. Amazon dealt with multiple outages driven by AI-assisted code changes, losing them millions of orders. Engineers are reporting that AI-generated code is producing technical debt at three to four times the previous rate.
The takeaway for leaders isn't to pump the brakes on AI in software development. It's that the organizations getting the most value are the ones investing in human oversight alongside AI speed.
TechCrunch: Are AI tokens the new signing bonus? →Bell's $1.7 billion, 300 MW facility in Saskatchewan will be the largest purpose-built AI data centre in Canada, with projected economic value of up to $12 billion for the province. It's part of Bell AI Fabric's sovereign compute network, which also includes a BC "super cluster." The Canadian AI infrastructure story from our last issue continues to accelerate.
Read more →We've been working extensively with Claude Cowork since its launch and are now using that experience to help clients with early testing of Copilot Cowork. If your organization is on M365 and interested in exploring what agentic AI can do inside your existing tools, reach us at hello@senseandmotion.co.
We're trying to build a real picture of how local organizations are using AI. We're interested in what tools are getting used, where you're seeing the most success, and where efforts are stalling.
We've put together a quick survey (less than five minutes) and if you complete it, we'll send you a $10 coffee gift card as a thank you. Your responses directly shape what we write about and how we advise our clients. The data from these surveys has consistently been some of the most valuable input we get.
Take the Survey →Thanks for reading the second issue of Making Sense. If you found something useful here, forward it to a colleague. And if you have suggestions for future content or want to talk about what AI could mean for your organization, drop us a line at hello@senseandmotion.co.
If this was forwarded to you, subscribe to future issues here.
Sense & Motion Inc. #200 – 410 W Georgia St, Vancouver, BC