I Built a SMART Goal Coach Gem — Here's What It Actually Does

Keywords: SMART goals, goal setting, AI for managers, Gemini Gem, OKRs, quarterly planning, performance goals, personal development

Reading time: ~7 minutes

Most goals I set for myself and my team start as intentions, not plans. "Improve response time" is an intention. "Reduce average incident response time from 45 minutes to under 20 minutes by Q3, measured by ticket system timestamps" is a goal. The difference between the two determines whether anyone actually makes progress — and whether you can tell if they did.

I've been managing teams for over two decades. The SMART framework — Specific, Measurable, Achievable, Relevant, Time-bound — has been a workhorse for most of that time. The challenge has never been knowing the acronym. It's been applying it rigorously under time pressure, in conversation with people who resist making commitments concrete, and on goals that resist easy measurement.

My SMART Goal Coach Gem doesn't write the goals for me. It asks the questions that turn vague intentions into trackable commitments — the questions I'd ask if I had an hour to coach each person one-on-one and the time to think carefully about their situation.

The Problem I Was Trying to Solve

Three specific patterns kept showing up in goal-setting conversations before I built this Gem:

First, goals that sounded good but couldn't be measured. "Improve documentation quality" is the kind of goal that everyone agrees with and no one can evaluate at year-end. Is documentation better? Compared to when? By what standard? The Gem asks the questions that turn it into something measurable — and if the person can't answer those questions, the goal probably wasn't ready to be written down.

Second, time bounds that didn't actually bind. "By Q3" sounds specific but if Q3 is six months away and the work is incremental, the goal allows too much drift. The Gem pushes for tighter timeframes where they're realistic, and surfaces the tradeoffs when they aren't.

Third, achievability discussions that didn't happen. Some goals are too aggressive and everyone knows it; some are too easy and are just filler. Neither serves the team. The Gem helps separate "stretch but possible" from "impossible" or "trivial" by asking for evidence the person has for their estimate.

I wanted a tool that ran the questions I'd run in a one-on-one conversation — but available instantly, with consistent rigor, even when I'm short on time.

What I Tried First (and Why It Wasn't Enough)

My first approach was a checklist I gave out at goal-setting time: "Is your goal specific? Measurable? Achievable? Relevant? Time-bound?" The intent was right but the checklist got handed back with check marks and the goals were still vague. Checking boxes is not the same as answering the underlying question.

Second attempt was a written template with fields to fill in. Slightly better — at least the fields prompted thinking — but people wrote "improve X" in every field instead of answering what "improve" actually meant for them.

Third attempt was having one-on-one conversations with each team member during the goal-setting window. The most effective by far. But I had 8–10 people on the team and two weeks to do it, and I couldn't give each conversation the time it deserved. The conversation approach worked; the time it took didn't scale.

What I wanted was a way to run a focused goal-setting conversation with each person — without requiring me to be present for all of it.

The SMART Goal Coach Gem did exactly that.

The Gem I Use

Here's the core of the instructions I saved into the Gem:

You are my SMART goal coach. You help convert intentions and vague goals into specific, measurable, achievable, relevant, and time-bound commitments.

When someone shares a goal they want to write down:

  • Ask the questions needed to test each SMART criterion — don't assume the answer
  • Push for measurable, not just "I'll know it when I see it"
  • Push for time-bound, not just "soon" or "by next year"
  • If achievability is unclear, ask what evidence they have for the estimate
  • If relevance is unclear, ask how this connects to a larger objective (team, company, personal)
  • Rewrite the goal in clear, committed language once it's specific enough
  • Flag any criterion that still doesn't pass after questioning

My context: engineering team manager in IT infrastructure. Goals are usually quarterly or annual, tied to performance reviews and team OKRs. I value rigor over encouragement.

The "I value rigor over encouragement" instruction is the one that surprised me by mattering most. The default tendency of any AI assistant is to encourage — "great goal, you can do it!" — which is the opposite of what goal-setting conversations actually need. The Gem asks harder questions because I've explicitly told it that's what I want.

Two Real Examples — Before and After

These are based on actual goal-setting conversations from the last few review cycles. Names and specific work are changed but the goal transformations are real.

Example 1: Vague Intention → Measurable Goal

Situation. A team member brought me this goal at the start of a quarter: "I want to get better at monitoring." Useful direction, not a goal. I needed to convert it into something trackable in three months.

What I asked the Gem:

"A team member wrote 'I want to get better at monitoring' as a quarterly goal. Make it SMART."

Gem's response:

This goal needs work on three criteria before it's ready to commit to. Questions I'd ask:

Specific: What aspect of monitoring? Alert tuning? Dashboard coverage? On-call response? Log analysis? Each is a different skill set.

Measurable: What would "better" look like as a number? Coverage gaps closed? MTTR reduction? False positive rate?

Time-bound: Q3 covers ~13 weeks. What's the intermediate milestone at week 6?

Suggested rewrite after those questions are answered:

"Reduce alert noise on the [specific system] monitoring by 30% by end of Q3, measured by alert volume per week, achieved by identifying the top 5 noise contributors and tuning or removing each by week 8."

What the Gem got right. It didn't write a goal for the team member — it wrote the questions I would need to ask to help them write their own. The suggested rewrite at the end was illustrative, not prescriptive — a model to react to. The questions were ordered by which criteria fail most often for goals like this (specific, then measurable, then time-bound), which matches what I'd ask in person.

What happened. The team member came back with a goal scoped to dashboard coverage (specific), with a target of three new dashboards deployed and two existing ones tuned (measurable), with a milestone at week 6 and final delivery at end of Q3 (time-bound). The Gem's questions did most of the work; I confirmed in a five-minute conversation.

Example 2: Aggressive Timeframe → Realistic Commitment

Situation. Another team member proposed migrating our entire log aggregation stack to a new platform in one quarter. That's a six-month project at minimum. The goal was achievable eventually but not in the timeframe proposed. I needed to either push back on the timeframe or break the goal into something realistic.

What I asked the Gem:

"Is 'migrate the entire log aggregation stack to new platform in Q3' achievable for one engineer alongside their other responsibilities?"

Gem's response:

Almost certainly not. A full log stack migration typically breaks into:

  1. Evaluation and architecture decision (2–3 weeks)
  2. Pilot deployment on one non-critical system (3–4 weeks)
  3. Migration of remaining systems in priority order (8–12 weeks)
  4. Decommission of old stack (2–3 weeks)

That's roughly 15–22 weeks for one engineer doing it alongside other responsibilities. Q3 alone (~13 weeks) is not enough.

Two realistic options for the goal:

Option A — Reshape the goal: "Complete evaluation, architecture decision, and pilot deployment on one non-critical system by end of Q3." (~7–8 weeks)

Option B — Extend the timeframe: Keep the full migration as the goal but stretch to two quarters (Q3 + Q4) with clear milestone gates between.

What the Gem got right. It broke down the work into the actual phases, gave a realistic time estimate based on those phases, and offered two concrete options rather than a vague "this might be too aggressive." The breakdown is the kind of analysis I'd do in my head; having it written out makes the conversation easier.

What happened. The team member chose Option A — a Q3 milestone that's measurable, achievable, and sets up the Q4 work. The original goal was ambitious but not achievable as written; the reshaped version is a real commitment.

Where This Works and Where It Doesn't

After about 18 months of use across two full review cycles, here's an honest assessment.

It works well for:

  • Vague intentions that need to be measurable. "I want to get better at X" → concrete goal with evidence of progress.
  • Time-bound problems. Goals that are too vague on timing or too ambitious for the timeframe.
  • Achievability pushback. When the proposed scope doesn't match the time available.
  • Self-coaching on personal goals. I use it for my own quarterly objectives with the same rigor I ask of the team.

It doesn't work well for:

  • Goals that depend on others' commitments. "Get engineering to reduce deploy time" is partly outside the goal-setter's control. The Gem can flag this but can't solve it.
  • Stretch goals that are intentionally ambitious. Some goals are meant to be barely-achievable by design. The Gem's achievability framing can flatten those.
  • Goals in areas where the person doesn't have clear metrics. "Become a better leader" is hard to make SMART even with good questioning — sometimes the metric isn't there yet.
  • Replacing the actual conversation. The Gem gives the questions and frameworks, but the goal commitment still needs the human conversation to land.

How to Use This With a Team

The workflow that worked for me:

  1. Ask each team member to draft 3–5 quarterly goals before our goal-setting conversation. No template required.
  2. Run each draft through the Gem. Get the questions and the suggested rewrite.
  3. In the conversation, use the Gem's questions as scaffolding — but the conversation itself is human. The Gem doesn't replace the discussion.
  4. After the conversation, the team member updates the goal in their own words based on what we discussed. They own the final wording.
  5. Review the goals quarterly against actual progress, using the same rigor the SMART framework established.

The most common mistake is treating the Gem's output as the final goal. It's not. It's the input to the conversation that produces the final goal. People who try to skip the conversation get goals that look SMART but don't reflect the team member's actual thinking — and they fall apart at the first unexpected obstacle.

A Note on What This Replaces

The Gem doesn't replace the manager's judgment about whether a goal is realistic given what they know about the person and the team. It doesn't replace the conversation where the commitment actually happens. And it doesn't replace the work of actually pursuing the goal over the quarter.

What it replaces is the awkward silence when someone says "I want to get better at monitoring" and you don't immediately know what questions to ask to make it concrete. The questions are now always available. The team member's answers are still their own. The manager's role in synthesizing those answers into a goal is still theirs. But the initial legwork — the rigorous questioning that turns intentions into commitments — is faster and more consistent.

For me, the Gem has cut goal-setting conversation prep time by roughly 60% per person. Over a team of 8, that's an entire day back at every review cycle.

Related Reading

Sources

— Justin

📅 First published: 2026-05-08 | 🔄 Last updated: 2026-06-21