Cohort Insight Report · AI for Leading Efficiently
AI for Leading Efficiently
An open badge that asked education leaders to pick one operational headache, talk it through with a partner, and try handing part of it to AI. This is what 18 people said across 36 sessions, read transcript by transcript.
How the cohort engaged
AI for Leading Efficiently ran as an open community badge, offered in partnership with the Association of Washington School Principals for Washington clock hours. Educators enrolled individually, among them some members of the Mark Cuban Foundation Teacher Fellows. The room was mixed: classroom and special education teachers alongside coaches, district and regional leaders, university faculty and administrators, and education consultants. Everything below is reported in aggregate.
Each dot is one session, in the order they ran. Session 1 was paired by design and Session 2 was a solo reflection. Sessions 3 and 4 let people choose, and 10 of those 15 runs were paired. People per session ran 18, 10, 12 and 12: four educators stopped after Session 1, and one went straight from Session 1 to Session 3.
When sessions ran. A paired session needs two calendars to line up, and the badge's rhythm shows it: a slow late-July start, nearly everything in August, and a short September close.
Four of the five September sessions were solo reflections run on the badge's final day. One Session 2 reflection that ran after the window closed is not counted.
Every Idea was generated by the platform and written back to the participant in the second person, including the one flagged as not AI-suggested, so none are treated as participant voice. Two were held back in review. A seventh Small Win, a camera test, is excluded.
Success Signs are observable evidence indicators from the AI for Deeper Learning Framework. These counts come from the signs people chose when writing their own Small Wins. Only three of the six wins carried one, so the numbers are small.
The signs on wins lean toward students, though most of this badge's work was adult-facing. Session 3 asked partners which signs they had seen or hoped to see, and those answers fit the badge more closely: time, accuracy, fewer errors and colleagues coming along.
The shape of the learning
Eighteen people each brought one operational problem and spent most of the badge talking it through with someone in a different job. The problems were ordinary and heavy. IEP drafting, weekly newsletters, survey analysis, a professional development calendar rebuilt from years of old agendas, and a downloads folder holding roughly 2,500 unnamed screenshots all came up. The record is unusually concrete about what happened next. People named the tool, described the steps, and in several cases put a number on the time.
By Session 4 the cohort had split along a line that had little to do with enthusiasm. A handful had a workflow in daily use, such as an IEP drafting agent shared across a team, a morning briefing prompt in its sixth version, and web apps that send a self-evaluation straight to a supervisor. Others finished with a clear plan and a plain account of why it was not running yet. Usually that was time to learn the tool, an approval still pending, or a feature switched off by IT. Both groups described the same tradeoff, which is that the hours come back only after hours go in.
The learning target asking leaders how operational demands shape their focus on equity drew the thinnest evidence. When equity came up, it was mostly about access to AI itself, such as who can afford the stronger model and which tools students are cleared to use. Few connected their own workload to the equity work it crowds out. That reads as a gap in the badge's prompts more than in the people taking it, and it is the clearest place a next version can push.
It is making my work easier, but it also making it harder because I'm getting more done.
I'm not trying to create extra work, but the work that I've already been doing, I ask it to translate it into a workflow.
it's maybe not operational, it's maybe not transformational, but it's like, relational.
Five themes that surfaced
Where a number appears, it counts distinct people out of the 18 with a recorded session. Counts come from reading the transcripts in context, with prompts read aloud and background audio set aside, so treat each one as a floor. Partners in a paired session are counted only for what they said about their own work.
At least six described the same arithmetic. A tool might save hours every week, but learning it takes hours this week, and every week was busy. Several named that trade out loud and kept going. Others admitted the pull to fall back on the familiar way under deadline.
At least eight began by loading a tool with material they already trusted, from an IEP writing library to a leadership framework, district priorities, curriculum standards and years of PD agendas. Output improved as sources grew. One drafted a team template 75 percent of the way in half an hour, then finished by hand.
At least seven tied efficiency to people: more conversations with the leaders they support, more time with students, a calmer team. Several also watched saved time refill with more work. One district leader argued the time matrix has no box for relational work, often the most important kind in schools.
At least seven were waiting on something outside their control, such as district approval for student data, a single sanctioned tool, corporate IT settings or a paid license. Much of that caution is well founded, since student records were often at stake. Where policy trailed practice, people paused or rebuilt their work to fit.
At least seven treated their own learning as a leadership move. They planned short AI check-ins at faculty meetings, invited new staff to shadow them, restarted a district AI committee whose members arrived far apart, and let a teammate struggle productively. Several named the discomfort of leading while still learning.
Themes were built from the 36 recorded sessions and the six hand-written Small Wins. Transcripts were read in full, session by session, then coded and checked back against the whole corpus. These are automatic voice transcriptions, and passages that were not a participant speaking for themselves were set aside before any coding or quoting: speech from other people in the room, stretches where recognition repeated one word or phrase dozens of times, garbled fragments, and prompts and resources read aloud. One session ended normally but captured no audio, and the platform-generated Ideas were treated as already-analyzed artifacts rather than as participant voice.
What this badge was built to develop
AI for Leading Efficiently carried three learning targets, written against PSEL Standard 9 and CCDEI Standard 4. The cards below read them against the closest competencies in the AI for Deeper Learning Framework, a mapping made for this report. This is where evidence from the badge should be read first. Evidence depth reflects the richness and specificity of the available record, not a performance grade or evaluation.
Experiment/try new things with AI tools.
The deepest record in the badge. Sixteen of the 18 described trying a specific tool on a specific task, and most named it: Copilot agents, Gemini gems, Claude projects, NotebookLM notebooks, interactive slide decks, and one planning add-in that would not cooperate. The experiments ran from renaming a folder of files to building dozens of small apps, and almost all of them began with a real deadline.
Adapt/iterate your practice & use of AI tools, based on your learning.
Session 3 asked what each person would carry into the next iteration, and the answers were specific. An agent that kept forgetting its corrections led to a dedicated review agent and a shared checklist. A morning briefing prompt reached its sixth version. A project that rebuilt a whole spreadsheet was cut back to a list of changes, because the manual step turned out to be just as fast. About ten of 18 showed a change made because an earlier attempt taught them something. Several were still on their first attempt when the badge closed, so this reads as timing more than reluctance.
Leaders develop the professional capacity of school personnel through job-embedded learning.
At least 12 of 18 described building other people's capacity alongside their own. New case managers shadowed a colleague's AI-supported IEP drafting, a review checklist went out to a whole team, faculty meetings made room for AI talk, and one consultant urged a school to grow its own AI champion. The equity half of this target is where the record thins. Equity was raised mostly as access to AI itself, and few connected their operational load to the equity work it displaces. That is worth designing for directly.
Explore the full framework map
All six strands and 33 competencies, showing where this cohort's evidence landed beyond the three the badge maps to.
These conversations touched competencies well outside the three above, which is worth seeing but not worth grading the badge against. Every rating below was set from a fresh reading of this badge's transcripts, competency by competency, counting distinct people with genuine evidence and discounting passages where someone was reading a prompt or resource aloud. No rating was carried over from an earlier badge.
Evidence depth reflects the richness and specificity of available evidence, not a performance grade or evaluation. Bold items mean documented evidence is present. Lighter items mean the pattern is emerging or not yet surfaced in this cohort's data. Professional Habits carries the record, as an operations badge for adults should, and the student-facing strands sit low for the same reason; LM1 Student Goals does not appear at all. PH3 Exchange is observed directly this time, including ideas that traveled from one conversation into a partner's later session.
sideby's AI for Deeper Learning Framework is informed by the UNESCO AI Competency Framework for Teachers, the colleague.ai AICE Framework, Teach AI with EC + OECD & code.org frameworks for AI learning, and the Hewlett Foundation Deeper Learning Competencies, evolving as AI evolves and shifts what this technology can mean for deeper learning and education.
Where evidence is strongest, session by session
This view reads across the whole framework rather than the three competencies the badge maps to, so it picks up everything the cohort touched. The four sessions mixed paired and solo formats, and each drew on a different part of the map. Evidence depth reflects the volume and specificity of transcript text recorded for each session, not a performance grade or evaluation.
Session counts come from the platform export and are distinct people, not runs. One Session 3 solo reflection ended normally but captured no audio. It is counted as a session but contributes no qualitative evidence.
What got written down
Six Small Wins were written by five people during the badge window, five logged as Successes and one as a Lesson Learned. They are the only place in this badge where someone chose to write something rather than answer a prompt. Private means visible only to the person who wrote it; launched means shared with the community, and two were. The six below are summarized rather than quoted, so the wording is ours and the substance is theirs.
While building a project in Claude, the author got stuck on how to word the instructions and asked the tool to write them in a form it could use well. The output improved and setup took far less time. Their recorded takeaway is that AI works as a thought partner on how to use AI, a move two other participants described independently in their sessions.
The same author found that very detailed project instructions made the tool excellent at one task and rigid at anything next to it. Their plan is to keep building task-specific projects with deliberate room for flexibility. It is the only Lesson Learned in the set and the clearest written example of iteration.
A special educator was awarded a mini grant to prototype an AI-supported system for tracking progress toward IEP goals, with a dashboard, simpler data collection and automated reporting. The stated aim is less data work for case managers and clearer, faster communication with families. The prototype is planned for spring, so this records a funded plan whose tool is still to come.
An educator paired AI analytics with a what, so what, now what reflection framework and used it to isolate and work on one specific weakness at a time. It matches the end-of-day reflection habit the same person described in a paired session.
An educator described a science student who attempts every lesson and assessment alone first, then books time to work through what did not land. Read against the framework's everyday indicators of deeper learning, his approach matched many of the learning skills and mindsets markers. AI does not appear in this win.
An educator turned a textbook business unit into a project where students found customer needs, built budgets and marketing plans, and pitched products. Students began asking whether customers would buy what they made, where before they asked what the teacher wanted.
All six real entries in the window are shown; a seventh, a camera test, is excluded. Small Wins carry no badge attribution in the platform, so these were filtered by person and date, and every author was active in another sideby badge during the same weeks. Two were written before their author's first session in this badge and two do not mention AI, so read the set as what these people were proud of that summer, which overlaps with this badge only in part.
Cohort voice across the badge
Every line below is exactly as a participant said it in one of the badge's paired or solo sessions, from a group that included classroom and special education teachers, coaches, district and regional leaders, university faculty and administrators, and education consultants. These are voice transcriptions, so the false starts and run-ons are theirs and are left as recorded. An open badge draws people already leaning in. The caution heard here comes from inside that group, and the colleagues some of them described as hesitant were not in these conversations to give their own reasons.
I feel like sometimes we can't move fast enough and yet at the same time I want to be slow.
AI is only good as the humans in the loop.
will it take me longer to figure out the app than it would to actually have done the task myself?
I used to spend hours on IEPs, and now I feel like, you know, under an hour I can complete everything.
it maintains an incredible level of perceived self-competence that I have to often correct for.
I know that's saving me time, but it's also creating more work.
you can choose not to use it, but you need to be informed. You need to know what you're not choosing.
so that they're able to use it as a learning tool and not just an answer factory.
I need to look at some of those tasks that are really inconvenient as transformational opportunities, right?
It's also made me a lot more cautious about AI, because I realized that I don't really understand what's going on behind the scenes
it's like we still i step in the air uh do you know what i'm saying it it's not step in the concrete yet
I need the sounding board. I need other people on a journey, whether they're at the same place or at a different place.
When they are a part of the planning process, they're more involved in the other aspect of utilizing it.
I just gave myself permission. And then what did I do? I got away from it.
I'm too busy. I'm doing too many things. And I'm not sure all of the things that I do every day are prioritized correctly.
It definitely feels like the technology is evolving with us as we use it
I want to thought partner with a human.
I really wanna try to model kind of my AI journey publicly
Because I believe in, like, when students do an assignment, within a week they need to know if they pass or fail, or if they master the standards
we know we put everything on teachers, but this is how you do this easier. And it's like, okay, but you're still putting more on your teachers.
From one badge to the bigger questions
AI for Leading Efficiently is a completed badge, and its record is useful well past the people who took it. What follows is what it demonstrated about open, peer-based learning around AI for deeper learning, where the badge itself goes next, and the questions it puts to the field.
By the final session, several participants had a workflow in daily use, including an IEP drafting agent most of a team now works from and a morning briefing prompt one leader said saves 15 to 45 minutes a day. The rest could say exactly what stood between them and a working version, which is a useful place to end a four-session badge.
Twenty-one paired conversations produced 21 different pairings. Classroom teachers talked with district leaders, and a university administrator learned from a faculty developer who builds apps. Ideas moved between people: a suggestion to try NotebookLM, a way to map a workflow before building anything, and an idea about precision learning that one participant carried into a later session with someone else.
One-word check-ins in Session 3 ran from excited to inundated, and people at both ends finished the badge. Participants who arrived unsure which tool to open left with one chosen and a next step named, often on a partner's suggestion.
The badge already sets aside at least 90 minutes to build and try a workflow between sessions, and each session now comes with a clearer preview. The next step is to point people to the resources that exist before, during and after every session. The badge deeper dive on the learning page holds the full set of information and resources, and the preview for each session on the match and session page shows the exact prompts and in-session resources ahead of time. Two participants found the example resources hard to take in during the three minutes a prompt allows, so a look ahead lets people arrive ready to talk.
In this cohort the hoped-for answer was people, meaning more conversations with staff and more time with students. Several also noticed the hour refilling with more work before anyone had chosen how to spend it.
Participants felt the pull to be the person who already knows. The stance they moved toward was a fellow explorer who keeps AI on every faculty meeting agenda until it no longer needs a slot.
Here the caution was often warranted, since student records were at stake. The cost showed up as tools rebuilt after the fact to fit policy, plans paused for approval, and students cleared to use no AI tool at all.