AI in OKR Execution: Why Most Agile Tools Still Can’t Close the Gap
Most conversations about AI in enterprise planning start with automation. How much faster can a report be generated? How many meetings can be summarized? How many manual handoffs can be eliminated?
These are useful questions. But they’re not the most important ones.
The deeper question — the one that determines whether AI actually improves how organizations deliver value — is this: does the AI understand why the work exists?
At the team level, that “why” is obvious. A developer picks up a task, sees the objective it supports, and understands the intent behind it. The context is right there — in the conversation, the backlog, the team’s shared understanding of what this quarter is supposed to accomplish.
But as organizations scale, that clarity erodes. Work gets handed off, translated, and decomposed across layers of planning until the original intent is buried somewhere in a hierarchy of epics and stories that no one has time to trace back. By the time a team is executing, the connection between their daily work and the organizational OKRs it was supposed to support has become theoretical at best.
The Gap Between Setting Objectives and Achieving Them
OKRs are the dominant framework for strategic alignment in scaling organizations and for good reason. They force clarity about what success looks like, create measurable outcomes at every level, and in theory, connect every team’s work to something that matters at the organizational level.
In practice, the gap between setting OKRs and actually executing against them is where most of the value gets lost.
Once teams pass the planning stage, they respond to what’s urgent rather than what’s strategically important. New work gets added without a clear trace back to existing objectives. A dependency shifts, and the work around it gets reorganized — but the OKR it was supporting doesn’t get updated to reflect that. By mid-quarter, there’s often a significant gap between the objectives that were committed to and the work that’s actually happening, and nobody has a clear picture of how large that gap is until it’s too late to close it.
This is the connection between objectives and execution where context is supposedly maintained manually, and at scale, that manual maintenance is too much to ask of the people who are also trying to deliver.
Difference between AI that identifies this gap and the one that doesn’t
The word AI gets attached to a lot of tools that aren’t doing anything genuinely contextual. A general-purpose assistant that answers questions based on a prompt is generating plausible text based on pattern matching. That’s useful for drafting an email or summarizing a meeting but it doesn’t understand your organization, your objectives, or the history of decisions that produced your current plan. Every time you use it, you’re starting from scratch.
Context in AI is different in one specific, meaningful way: it operates from within the system where the work actually lives.
Get Kendis - Streamlining Agile at Scale’s stories in your inbox
Join Medium for free to get updates from this writer.
Instead of asking you to explain the situation in a prompt, a contextual AI reads the artifacts that already exist; the objectives your teams committed to, the work items linked to them, the way priorities have shifted over time, and uses that grounding to generate analysis that’s relevant to your specific situation. The intelligence isn’t in the output alone. It’s in what the AI is reading before it responds.
That distinction matters enormously once you’re managing OKRs across dozens of teams and themes. The gap between knowing what is happening and understanding why it matters and whether it’s still connected to the goals that were set — is exactly where most planning intelligence breaks down.
How Kendis AI Preserves the “Why”
Kendis has built its AI capability-Charlie, directly into planning and execution workflow, not as a bolt-on assistant but as a context-aware layer that reads from the actual planning data inside Kendis.
The distinction is important. Charlie doesn’t ask you to describe your data. It already knows your data.
AI-Powered PI Objectives
One of the most visible failure points in PI or Quarterly planning is the objective-writing process. Teams spend hours discussing scope, negotiating dependencies, and making trade-offs. Then, in the final minutes of the planning event, they’re asked to compress all of that into clear, shared objectives — often while exhausted and under time pressure.
Kendis AI looks at what already exists: features, teams, backlog items, and the conversations that shaped them. From that context, it creates PI objectives that reflect the work teams have actually committed to — not a polished summary written from memory. The output gives you a starting point that already understands scope, intent, and outcomes.
Importantly, teams can choose whether objectives are compact and outcome-focused or more detailed, carrying context for stakeholders who weren’t in the room. The same planning inputs produce different output depending on the audience. That flexibility is only possible because Charlie understands the underlying plan — not just the words used to describe it.
OKRs and Strategic Themes: The Alignment Layer
Context-preservation becomes most critical when you trace the chain from team-level work all the way up to organizational strategy.
Kendis supports Strategic Themes and OKR Management by aligning high-level strategic themes with operational goals through robust OKR (Objectives and Key Results) management capabilities. But alignment on a dashboard and alignment in practice are different things. Kendis AI closes that gap.
When an AI can read both the Strategic Themes set at the portfolio level and the features committed by teams in the current quarter, it can identify misalignment before it becomes a delivery problem. A feature that doesn’t trace to any active Strategic Theme is a planning gap worth flagging. An OKR with no committed features backing it is a strategy commitment that may not be fulfilled.
This kind of contextual analysis isn’t possible with generic AI tools. It requires an AI that understands the full hierarchy — from the organizational vision down to the sprint-level task — and can surface the points where that chain breaks.
