Tecnet’s perspective on the Strategic CIO’s Generative AI Playbook — and what we’re actually seeing in BC organizations
Microsoft’s Strategic CIO’s Generative AI Playbook is a good document, and we agree with almost all of it. Its core argument is one we’ve been making to clients for two years: AI is not a technology project. It’s a business transformation that happens to involve technology, and it fails when it’s handed to IT and treated as a rollout.
Microsoft’s newly released 2026 Work Trend Index makes the same case with a year of fresh evidence behind it, and sharpens it considerably.
But most of the organizations we work with across British Columbia don’t have a CIO. They have an owner, a CFO, an operations lead, and maybe an IT manager who is already carrying too much. When they read a playbook written for enterprise CIOs, the natural reaction is: this isn’t for us.
It is. The principles hold at 60 users just as well as at 6,000. What changes is the scale of the work, not the sequence.
Here’s what the playbook gets right, and what we see when those ideas meet a real 50-to-250-person organization.
The problem: adoption is outpacing readiness
Microsoft’s 2026 Work Trend Index, released in May, surveyed 20,000 knowledge workers who already use AI at work across 10 markets. Its central finding has a name: the Transformation Paradox. Employees are ready to change how they work, yet the organizations around them aren’t built to capture it.
Only 19% of AI users sit in what Microsoft calls the Frontier zone, where individual capability and organizational readiness reinforce each other. Another 10% are actively blocked: capable people in companies that haven’t caught up. Half sit in the middle, with both still taking shape. And just 26% say their leadership is clearly and consistently aligned on AI.
Usage points the same direction. Active agents across the Microsoft 365 ecosystem grew 15-fold year over year, rising to 18-fold in large enterprises. AI is no longer something being evaluated. It’s in the building.
The results, however, are uneven. Gartner reports that 47% of CIOs say AI has not met their ROI expectations, and only 35% of data and analytics leaders can effectively demonstrate measurable value to stakeholders. Gartner also projects that 30% of generative AI projects will be abandoned after proof of concept; most often because of poor data quality, weak risk controls, or unclear business value.
That gap between adoption and outcome is the single most important thing in both documents, and it matches what we see. The organizations getting real returns from Copilot aren’t the ones who bought the most licences. They’re the ones whose environment was ready to give AI good answers.
The insight: AI exposes the state of your information
This is the part we’d underline hardest for any business owner considering AI.
Copilot doesn’t create new access to your data. It surfaces what a user could already reach. instantly, in plain language, without them having to know where to look. If your permissions are loose, that’s not an AI problem. It’s an existing problem that AI makes visible on day one.
The playbook calls this “oversharing” and recommends access reviews, sensitivity labels, and site ownership. In practice, for a mid-sized organization, it usually looks less like a governance program and more like a cleanup.
We worked with a private company of just over 50 users whose SharePoint environment had grown organically for years. Data was fragmented across sites and libraries, permissions were inconsistent, and IT had limited visibility into who could access what. Users struggled to find files; the business was carrying compliance risk and redundant storage cost without knowing it. We rebuilt the structure: unique permissions on critical folders, security groups to replace one-off access grants, a controlled migration to remove redundancy, and training so people could actually navigate the result.
None of that was an AI project. But it’s precisely the work that determines whether an AI tool returns something useful or something wrong, and whether it quietly surfaces a salary file to someone who shouldn’t see it.
The pattern repeats. A property management company we migrated to Microsoft 365 had over 230 user mailboxes, 120 shared mailboxes, and 21 public folders; more than 300GB of accumulated data on a legacy system with no modern security controls. Getting to a place where AI could add value meant first getting to a place where the data was governed at all. We ran the migration in phases with zero data loss, then rolled out MFA through self-enrolment followed by conditional enrolment, reaching 96% compliance.
The 2026 Work Trend Index puts a number on why this matters. Across 29 factors tested, organizational conditions — culture, manager support, talent practices — accounted for roughly twice the AI impact of individual mindset and behaviour (67% versus 32%). The strongest single predictor of whether someone gets real value from AI wasn’t their skill with it. It was the environment around them.
The insight is simple: AI readiness and IT hygiene are the same project. Most organizations we meet are further from “AI-ready” than they think, and closer than they fear, because the work required is work they already needed to do.
The framework: four questions before you buy a licence
The playbook recommends CIOs prioritize AI investments against business impact, feasibility, and quantifiable return. We use a simplified version with clients who don’t have an AI council or a data team.
1. What specific work are we trying to make faster? Not “we want to use AI.” Something a person does on a Tuesday. The playbook’s Access Holdings example is instructive: report preparation dropped from six hours to 45 minutes. That’s a measurable claim about a specific task. Start with two or three of those, usually in finance reporting, HR administration, marketing content, or proposal writing.
2. Can our environment actually support it? Are files where they should be? Are permissions current? Is old, superseded content archived, or is a five-year-old price list still sitting in a live library waiting to be quoted back to you? AI has no sense of what’s stale.
3. Who owns the outcome? The playbook is emphatic that AI adoption fails when it’s owned solely by technology teams; Gartner found 27% of chief data and analytics officers cite lack of business stakeholder involvement as their single biggest challenge. In a smaller organization this is easier, not harder. The person who owns the process should own the AI applied to it.
4. How will we know it worked? Microsoft’s measurement model; readiness, adoption, impact, is sound and scales down well. Readiness: is the environment and the team prepared? Adoption: is anyone actually using it after week three? Impact: did the hours, the cost, or the customer experience move?
If you can’t answer question four before you start, you’ll be in the 47% who can’t demonstrate ROI afterward.
The solution: people, then process, then tools
The playbook notes that 69% of CIOs plan to upskill employees on AI, but only 15% of IT leaders believe their workforce is genuinely prepared. That gap doesn’t close by buying licences.
We see the same thing on every deployment, AI or otherwise. When we rolled out 93 preconfigured Macs for a client, the technical win was cutting setup from two hours per device to 30 minutes. But the reason 90% of users reported a smooth transition was the unglamorous part: preference surveys before ordering, clear instructions in the box, and support available when someone got stuck. Technology adoption is a people exercise with a technical component, not the reverse.
The same holds for AI. Short, role-based training beats a comprehensive course nobody finishes. A safe space to experiment beats a policy document nobody reads. And a light governance framework, what data can go into AI tools, who reviews AI-generated output before it leaves the building.. beats an enterprise governance program you’ll never staff.
One more thing the playbook doesn’t dwell on, and we will: proactive beats reactive. For a non-profit with over 300 staff delivering essential services across BC, visibility into asset and performance data through our client portal let us identify a failing server platform before it caused an outage, not after. That same discipline of knowing the state of your environment before it forces a decision, is what separates organizations that adopt AI deliberately from those that discover it’s already in use and try to catch up.
Where to start
Microsoft’s playbook closes by asking whether you’re ready to shape the next era of digital transformation. It’s a fair question, but for most BC organizations it’s the second one. The first is smaller and more useful:
Do we know what’s in our environment, who can see it, and whether it’s current?
If the answer is yes, you’re closer to real AI value than most.
If it’s not, that’s your AI project.. and it’s one that pays for itself whether or not you deploy a single Copilot licence.
Download the PDF for the The Strategic CIO’s Generative AI Playbook
Not sure where AI fits into your organization?
Tecnet’s AI Discovery Session is a guided consultation that assesses your readiness across strategy, data, workforce, and infrastructure, identifies quick wins with measurable value, and builds a secure, responsible roadmap all scaled to your organization and aligned with BC privacy standards.
Learn more about our AI Discovery Sessions
Sources
- The Strategic CIO’s Generative AI Playbook, Microsoft, 2025
- Agents, human agency, and the opportunity for every organization, Microsoft 2026 Work Trend Index Annual Report, May 2026
- Gartner, 2024 CIO Survey
- Gartner, How to Calculate Business Value and Cost for GenAI Use Cases
- Gartner, The CIO Report: Gartner Answers to Top CIO Questions
- Tecnet client case studies (SharePoint optimization, Microsoft 365 migration, device deployment, non-profit infrastructure modernization)