How Our AI Coach Turns Feedback Into a Growth Plan

You Got Your Feedback. Now What?

You just opened your OfficePoll report. Six category scores, a narrative synthesis, honest observations from the people you work with every day. It is the clearest mirror you have ever had at work.

And then nothing happens.

This is the "so what" problem, and it is surprisingly common. A landmark meta-analysis by Kluger and DeNisi, covering 607 effect sizes across 131 studies, found that feedback improves performance only about 60% of the time. In the other 40% of cases, it had no effect or actually made things worse.

How can honest, accurate feedback backfire? Because without structure, feedback pushes people into identity-level thinking — "What does this say about me as a person?" — instead of task-level thinking — "What should I do differently next Tuesday?" That shift from task to self is what kills progress.

A separate meta-analysis by Smither, London, and Reilly confirmed the pattern with 360-degree feedback programs specifically: improvement is significantly larger when feedback is accompanied by coaching. Feedback creates awareness. Coaching creates action.

That is why we built the AI coach. Not because feedback does not matter, but because feedback alone is not enough.

The Framework Behind Every Conversation

The OfficePoll coach uses the GROW model, the most widely adopted coaching framework in the world. Developed by Sir John Whitmore in the late 1980s and used by organizations from Google to the NHS, GROW gives every coaching conversation a spine. The acronym stands for four phases:

• Goal — What do you actually want to change? Not what "should" you change. What matters to you?
• Reality — What does the data say? This is where your feedback report lives. The coach helps you sit with the numbers before jumping to solutions.
• Options — What could you try? The coach expands your thinking beyond your default patterns, brainstorming without judgment.
• Will — What will you do? Not "should." Will. A specific commitment with a timeline.

Without this structure, coaching conversations drift. Someone vents about a low score, feels a little better, and walks away without changing anything. GROW prevents that by making sure every conversation moves from understanding to action.

Each conversation with the OfficePoll coach spans about 8 to 12 exchanges and takes roughly 10 minutes. A progress bar at the top of the screen shows exactly where you are: Goal, Reality, Options, or Commit. There is no mystery about what comes next.

It Starts With Your Data, Not a Blank Page

One of the most common complaints about coaching tools is the dreaded blank-page opening: "What would you like to work on today?" That puts the burden on you to diagnose your own development needs, which is exactly the kind of thing most people are not great at doing objectively.

The OfficePoll coach does not do that. When you open your report, the coach has already analyzed your scores and narrative synthesis to identify three or four topics worth exploring. These fall into categories:

• Growth areas — categories where your scores suggest room for improvement
• Blind spots — gaps between how you see yourself and how others experience you
• Strengths to leverage — high scores that could be applied more deliberately
• Continuation topics — in subsequent cycles, areas where you made commitments last time

You pick a topic. The coach opens with specifics:

"Your mentorship score is 2.4, your lowest category. The narrative synthesis mentions that junior colleagues feel they don't get enough context on decisions. Let's talk about what a realistic improvement would look like for you."

That is a very different starting point than "Tell me about your development goals." The coach leads with substance because the research is clear: coaching grounded in concrete data produces better outcomes than coaching based on self-report alone.

The Commitment That Actually Sticks

Every coaching conversation ends in the Will phase, where you create what psychologists call an "implementation intention" — or more simply, an if-then plan.

This is the single most research-backed technique for converting goals into action. A meta-analysis by Gollwitzer and Sheeran covering 94 independent studies found that if-then plans roughly double the likelihood of achieving a goal compared to goal intentions alone, with a medium-to-large effect size.

Why do they work? An if-then plan creates a mental link between a situational cue and a response. Instead of relying on willpower ("I should communicate more proactively"), the situation does the work for you:

"If I finish a major decision, then I will send a brief update to affected stakeholders within 24 hours."

The cue triggers the behavior, so you are not constantly spending mental energy remembering to do the thing. The research calls this "automating goal-directed responses." In plain language, it means your environment starts doing some of the work for you.

Here is another example from an actual coaching conversation about mentorship:

"If a junior colleague asks for help, then I will ask what they have tried first and guide them to the answer rather than giving it directly."

Notice what these commitments have in common: a specific trigger, a specific behavior, and no room for ambiguity. The coach helps you refine your if-then plan until it hits that standard. Vague commitments like "be a better mentor" do not make the cut.

The 90-Day Rhythm

Personal development is not a one-time event. The OfficePoll coach operates on 90-day coaching cycles, and there is real science behind that number.

The popular claim that it takes 21 days to form a habit is a myth. Research from University College London found it takes an average of 66 days for a new behavior to become automatic, with wide individual variation. A 90-day window gives you a realistic runway.

Quarterly cycles also align with how high-performing organizations already think about goals. The OKR framework, popularized by Google and now standard across the tech industry, typically operates on quarterly cadences. It is long enough to accomplish something meaningful but short enough to maintain focus and allow course correction.

Here is what a 90-day cycle looks like in practice:

• Week 1: You receive your feedback report and work through coaching conversations on your top topics, creating if-then commitments for each.
• Weeks 2 through 6: You practice your commitments in your daily work. Your development plan is visible on your dashboard as a reminder.
• Mid-cycle check-in: Around the halfway point, the coach checks in on each commitment. You self-report whether you are practicing regularly, have tried a few times, or have not started yet. The coach responds with personalized encouragement or helps you adjust the plan if something is not working.
• Weeks 7 through 12: Continue practicing. The commitments that feel natural are becoming habits. The ones that are not working get adjusted.
• New cycle: Fresh feedback arrives. You can see what changed in your scores and narrative. New topics emerge. The cycle repeats.

This is how real improvement compounds. Not through a single insight, but through repeated cycles of feedback, reflection, action, and measurement.

What AI Coaching Can and Cannot Do

We are not going to pretend this is the same as working with a great human coach. It is not. Here is an honest assessment.

What the AI coach does well:

• Structured, data-grounded conversations. Turning feedback scores and synthesis into specific action plans is an analytical task, and AI excels at it. A 2025 Conference Board study found that AI can handle up to 90% of day-to-day coaching functions, with 96% of users reporting that the coaching felt personalized to their situation.
• Consistency and availability. The coach is available whenever you are ready to think about your development. No scheduling, no cost per session, no waiting list. This matters because traditionally, structured coaching has been reserved for executives with large budgets.
• No judgment. Some people find it easier to be honest with an AI about their weaknesses than with a human. There is no relationship to manage, no impression to maintain. You can say "I genuinely do not know how to give direct feedback to someone" without worrying about how it looks.
• Accountability without pressure. The mid-cycle check-ins create gentle accountability. The coach remembers your commitments and asks about them, but it does not guilt-trip you. If something is not working, it helps you adjust rather than doubling down.

What the AI coach does not do:

• Emotional processing. If your feedback triggers a strong emotional reaction — and sometimes it will — the coach is not the right tool for working through those feelings. It can help you process the data, but it cannot provide genuine empathy. It is transparent about this and will never claim to be human.
• Organizational politics. The coach does not know your company's power dynamics, unwritten rules, or political landscape. For navigating sensitive situations like managing up or handling a difficult peer relationship, a human coach or trusted mentor is more appropriate.
• Values-based decisions. Questions like "Should I stay at this company?" or "Is this role aligned with who I want to become?" require a kind of holistic understanding that AI does not have. The coach stays in its lane: turning feedback into behavioral change plans.

The Conference Board explicitly recommends what they call a "tiered and blended model" where AI handles the structured coaching work and humans handle the emotionally complex situations. We agree. The OfficePoll coach is designed for the 90% of coaching that is analytical and actionable. For the other 10%, we encourage you to seek out a human you trust.

Your Report Is Not a Grade. It Is a Starting Point.

The most important shift the coach helps you make is this: your feedback report is not a verdict. It is not a performance rating. Nobody is grading you.

It is a snapshot of how your colleagues experience working with you right now. Some of what you read will feel accurate. Some will surprise you. Some might sting. All of it is useful, but only if you do something with it.

The coach exists to help you answer the most important question your feedback raises: "Now what?"

It walks you through a proven framework. It helps you sit with the data before rushing to fix things. It pushes you toward specific, research-backed commitments rather than vague intentions. And it checks in to make sure those commitments do not quietly fade away.

Eighty-nine percent of people who use AI coaching say their session resulted in specific, actionable next steps. We think the other 11% just need one more exchange in the Options phase.

Your feedback is waiting. The coach is ready when you are.