For leaders, it also brings practical improvements to Team, Site, rubric, and subscription management. Across mobile and M2, reliability updates help active sessions continue with fewer interruptions.

Takeaways that build across activities

Professional growth rarely happens in a single session. It develops as educators reflect, try something new, and return with more evidence.

Now, Takeaways are cumulative across activities in the same Group. A teacher’s earlier reflections and feedback trends carry into later Takeaways, making it easier to see what is changing over time and where a pattern may deserve more attention.

Instead of treating every activity as a new starting point, M2 helps educators connect their feedback and reflection day to day.

Assign multiple Reflections in a Group

Teachers can now keep Reflection assignments open across multiple activities in the same Group. 

They can also share separate Reflection assignments through their learning management system (LMS), while participants can still open and complete earlier assignments.

This gives teachers more flexibility to build reflection into an ongoing sequence of learning without students losing access to prior work.

Improved transcript labels

Transcripts now make it easier to distinguish teachers from other participants. Clearer speaker labels give educators better context when reviewing a conversation and make feedback and quantitative metrics easier to interpret. 

Teachers can spend less time figuring out who said what and more time connecting evidence to the moments that matter.

Other improvements across M2

M2 6.1 also includes several improvements across the service:

• Clearer feedback and reflection: Improved video previews, transcript downloads, evidence display, and ownership controls make activities easier to revisit.
• More flexible rubric management: Leaders can sort and delete custom rubrics, receive updated premade templates, and view both active and inactive rubrics.
• Background audio sessions: Active audio sessions can continue when the mobile app is in the background, while observation video stops when the app is hidden or the screen is locked.
• Automatic battery protection: Recordings stop automatically at 5% battery and cannot begin below that level.
• Clearer Ask M2 controls: Cancel replaces Mute during playback, and repeated remote actions recover more reliably.

A Better Experience for Daily Use

As the school year begins, many educators are turning to M2 for every support, whether it’s to see far more feedback than ever possible, get a detailed record of their daily teaching, or elevate student thinking through reflection.

In the same way that feedback and reflection have compounding benefits for educators and students, the benefits of this release compound the more times you turn on M2. 

The post M2 6.1: Connected takeaways, better transcripts, and more flexible assignments appeared first on Swivl.

District leaders often need to balance two goals: aligning instruction to district priorities while giving schools and teachers support that reflects their specific needs.

M2’s rubric structure makes this easy by allowing you to layer criteria at three levels:

• Team: District strategic priorities
• Site: School-specific priorities
• Group Type (coming soon): Targeted initiatives or teacher groups

The key is knowing what belongs where, so everyone gets the feedback and scoring they need.

Team: Start with district strategy

When you add rubric criteria at the Team level, everyone in your organization will get feedback and scoring in these areas when they use M2. This means that the Team level is for the priorities you want reflected across the district. 

Start with your district strategic plan and identify the goals that translate into observable classroom practices. For example, if your strategic plan prioritizes:

• Student engagement
• Literacy
• Critical thinking
• Instructional rigor
• Student-centered learning

Build criteria around what those priorities look like in practice.

You can also align M2 to your teacher evaluation framework and use an Evaluation group type to give teachers opportunities to practice and reflect on the same criteria used during formal observations.

Site: Add school-specific priorities

Use the Site level to build on the district foundation with priorities specific to a school.

Examples:

• An elementary school focusing on transitions and classroom routines
• A middle school emphasizing content-area literacy
• A high school working on formative assessment
• A school implementing PBIS or a new instructional model

Site-level criteria layer onto the Team baseline, so schools can address their own goals without losing district alignment.

Group Type: Target specific teachers or initiatives

Use a custom Group Type when the priority applies to a defined group rather than an entire school. 

Common uses include:

New teacher cohorts: Focus on classroom management, lesson structure, routines, engagement, or other induction goals.

Curriculum implementation: Create criteria specific to the instructional practices associated with a new curriculum or program.

Professional learning initiatives: Give a cohort of teachers a shared rubric tied to the skills they’re developing.

Time-bound initiatives: Keep temporary priorities targeted to the teachers participating in them rather than adding them to everyone’s rubric.

PLC collaborative planning: Create a group for teachers participating in a PLC and use criteria focused on effective collaborative planning, such as using student data, aligning instruction to standards, developing common assessments, or planning differentiated instruction.

Keep rubrics focused

Once your rubrics are built, look at the full teacher experience. Remember that each rubric will generate feedback and scores for teachers after an M2 teaching activity.

Avoid duplicating criteria at multiple levels or adding every district initiative to the Team rubric. Teachers should be able to quickly understand what they’re being asked to focus on and why.

A simple way to organize your rubrics:

Support teachers without overwhelm

A strong M2 rubric structure connects district strategy to school priorities to targeted teacher support.

Start with the priorities that matter across the district. Add school-specific needs at the Site level, and use Group Types when a smaller group needs more targeted feedback.

This gives leaders a scalable way to personalize support while keeping everyone moving in the same direction.

The post How to scale personalized support across your district: Best practices for M2 Rubrics appeared first on Swivl.

M2 6.0 is now available across web, mobile, and the M2 device. This release helps M2 better understand the work happening in each Group, whether it is a classroom, PLC, training, staff meeting, 1-on-1, or evaluation and provide more relevant support before, during, and after each session.

It also makes reflection easier to complete, brings mobile video recording to every user, and gives leaders clearer insights into instructional practice across their Teams and Sites.

Support that adapts

Different learning environments require different kinds of support. A classroom does not operate like a staff meeting, and a PLC should not be evaluated in the same way as a 1-on-1 conversation.

With M2 6.0, educators and leaders can select from six New Group types when creating new groups:

• Classroom
• Training
• Staff meeting
• PLC
• 1-on-1
• Evaluation

From the Admin Panel, Leaders and Coaches can create rubrics that can be applied to specific Group types, and the Tips users receive adjust based on the setting applied. The Evaluation Group type also creates a dedicated space for evaluation-focused work, either from Leaders or for Teachers practicing for formal observations. 

This added context helps M2 provide support that is more relevant to the purpose of each conversation or learning environment.

Takeaways that show shared understanding

Takeaways now bring together what developed across the full learning experience: the activity that occurred, the user’s Personal Reflection, and student thinking.

Instead of treating each source as a separate piece of information, Takeaways show the shared understanding that emerged across them. Educators can see what participants understood, connect that evidence to their own reflection, and identify clearer next steps for instruction.

Stronger insights for leaders and coaches

M2 6.0 gives Leaders and Coaches more ways to understand instructional practice without reducing it to a single score.

Teams can customize rubrics at the Team or Site level, making feedback more relevant to local priorities. Team Admin insights now include quantitative metrics such as talk time and question counts alongside Reflection Depth of Knowledge insights.

Together, these signals help leaders recognize patterns, understand how educators are engaging with their practice, and identify where additional support may be useful.

A lighter approach to Personal Reflection

Reflection should help educators make sense of their instruction, not feel like another task to complete.

Personal Reflections now use shorter, more focused questions based on the experience. An updated Stop/Continue flow also lets educators choose how many questions they want to answer, giving them more control over the depth and length of each reflection.

The result is a lighter experience that educators can complete while the experience is still fresh.

Reflections return to mobile devices

Participant Reflection Assignments and Personal Reflections are once again available on iOS and Android.

Teachers and students can now complete reflections from the device that is most convenient for them, making it easier to capture thinking in the moment rather than waiting until they return to a computer.

Record observation video from any mobile device

Educators can now record video directly from an iOS or Android phone or tablet, without connecting additional hardware.

Improvements across the experience made using recording more reliable from start to finish. This gives more users a simple way to capture evidence of their instruction wherever learning happens.

Find video faster

When video is available, My Feedback cards now display a video thumbnail directly on the cover.

This small change makes recorded sessions easier to recognize and revisit, helping teachers move more quickly between feedback and the moments that produced it.

Why this release matters

The most useful feedback begins with context.

By understanding the type of work happening in each Group and connecting what happened during an activity with reflection and student thinking, M2 can provide support that is more relevant to educators and more meaningful to leaders.

M2 6.0 brings that context together while making it easier for everyone to record, reflect, and turn everyday learning experiences into evidence for growth.

The post M2 6.0: More context, clearer takeaways, and easier reflection appeared first on Swivl.

School and district leaders constantly balance the big picture with on the ground details.

Across the organization, it’s important to get everyone working towards higher quality instruction, and to provide feedback and support that actually reflects that vision.

At the same time, you have a list of specific projects and cohorts that need extra attention: early career teachers, a curriculum investment, or a school that needs special support. These groups require customized feedback to match those goals, without forgetting the org-level priorities.

That is exactly what M2’s Sites within a Team can do.

Three levels that match how districts already work

An M2 account is organized in three levels.

The Team is your district. It holds organization-wide expectations: the Coherence Prompt and the Rubrics that describe what good teaching looks like everywhere.

A Site is a school. Every school in the district is its own Site inside the Team, with its own Coherence Prompt, its own custom Rubrics, and its own view of the data.

Groups live inside a Site and organize the work that happens below the building level: PLCs, grade levels, new teacher cohorts, curriculum implementations, leadership teams.

The same structure applies whether you have sixty schools or one. That consistency is what lets a district track progress, helps a principal see their own staff, and allows a coach to look at a single cohort.

Support org-level goals and offer customized support

Administrators deploy feedback Coherence Prompts and scoring Rubrics at two levels: Team and Site.

At the Team level, district leaders set the defaults. Everyone in the district works within that context. Teachers get more frequent, criteria-aligned feedback to improve their practice. Leaders get a consistent picture of how instruction is changing over time, across every building.

At the Site level, a school adds what matters in that building. The teachers there still belong to the district Team, and the feedback and rubrics from both levels work together to support their practice. For example, a district-wide focus on student discourse stays in place, while a middle school in its second year of a new literacy curriculum can add look-fors tied to that program.

With this setup, everyone gets targeted, more frequent feedback, while the schools with distinct needs get scoring and feedback to match their goals.

Groups hold the projects inside a school

Lots of work does not happen at the level of a whole building. It happens in a first-year teacher cohort, a fourth-grade team, or a department piloting new materials. Groups are how those cohorts live inside a school Site.

Each Group is created with a group type, which is how you segment the work in a building and later look at data for that segment. M2 includes six predefined types:

• Classroom
• Evaluation
• PLC
• Staff Meeting
• Training
• One-on-Ones

Additional group types are coming, so accounts will be able to define their own labels for things like a beginning teacher program or a building leadership team.

This is also how a district initiative that spans multiple schools works. An induction program or a curriculum pilot running in five buildings becomes a Group with members in any of those Sites. The shared expectations for that initiative can sit at the Team level, so every participating teacher hears the same thing, while each school’s Site prompt and rubrics keep the feedback grounded in that building.

Start with the work you’re already doing

You do not need a complete map of every school, cohort, and PLC before anyone benefits from M2 in your organization.

Set up the district Team, create Sites for the schools that are participating now, and build your first Groups around the work already underway. Identify the teachers you are giving the most support to, or the initiative taking up the most attention this year, and start there.

The rest of the district, whether that includes administrators, school-level leaders, or teachers, can keep working with M2 at the Team and Site levels. They still get feedback and scoring aligned to the organization’s high-level goals.

Coaches can work across Sites

Instructional coaches often do not belong to a single building. A coach supporting several schools can work across multiple Sites, which adds flexibility to your setup.

The teachers most closely tied to a building stay on that building’s Site. The coach does not have to be limited to one of them. They can move across the Sites they support, while the teachers in each Site still receive the prompt, rubrics, and data that belong to their school.

Coherence and Customization

When a district uses this structure, there are two important benefits.

Teachers across the organization get more frequent feedback than walk-throughs and formal observations ever can provide, and all of it is aligned to the district’s values. That is the promise of M2’s rubric scoring and Coherence Prompt: frequent, consistent, criteria-aligned support at scale.

Additionally, the schools and cohorts that need something more specific can also get that support every time they turn on M2. New teachers hear from M2 in language that matches their induction program. A PLC works on the practices they are trying to build together. A school gets data that reflects their teachers and their goals, not an average of the whole district.

Leaders do not have to choose between coherence and customization. M2 delivers both.

The post How M2 delivers coherence and customization: Sites within a team appeared first on Swivl.

Every great lesson starts with intention. Teachers plan with clear goals, meaningful activities, pacing decisions, instructional strategies, and the needs of their students in mind. But once teaching begins, even the best lesson plan can be hard to hold onto in real time.

That is why we built lesson plan upload for M2.

With this new feature, teachers can upload a lesson plan before they teach with M2. From there, M2 can use the plan as context to provide live tips and suggestions during instruction, then generate feedback after the lesson that is directly aligned to what the teacher intended to teach.

This is an important step toward what schools need most: coherence. M2 is designed to help bring curriculum, teaching, and school priorities together into one connected experience, so teachers are not receiving generic AI support or disconnected feedback. Instead, they receive guidance that reflects the lesson, the instructional goals, and the work actually happening in the classroom.

Why we built it

In many classrooms, coaching, feedback, curriculum implementation, and instructional improvement can feel like separate efforts. Lesson plan upload helps close that gap.

By allowing M2 to understand the lesson before instruction begins, the support becomes more relevant, more specific, and more connected to the teacher’s actual goals. M2 can help teachers stay aligned to the lesson they planned, reflect on whether key content was covered, and identify opportunities to strengthen instruction based on the strategies already embedded in the plan.

For school and district leaders, this also supports a more coherent instructional experience. When teachers use M2 with lesson plans, feedback can connect more clearly to curriculum expectations, instructional priorities, and classroom evidence. That creates better coaching conversations, more actionable reflection, and a clearer view of how planned instruction translates into practice.

How it works

Using the feature is simple:

1. Upload your lesson plan before teaching with M2.
2. Teach your lesson with M2 or M2 App active in the classroom.
3. Receive live tips and suggestions aligned to the lesson plan.
4. Review feedback after teaching that connects directly to the lesson goals, pacing, activities, and instructional strategies.

During the lesson, M2 can use the uploaded plan to provide more relevant in-the-moment support. After the lesson, M2 can generate feedback that helps the teacher understand how the lesson unfolded compared to what was planned.

What’s in it for teachers?

When M2 has access to the lesson plan, the feedback becomes grounded in the actual instructional context. They are receiving feedback tied to the lesson they prepared and the learning experience they delivered.

That can help teachers answer questions like:

• Did I keep pace with the lesson?
• Did I cover the key material I planned to teach?
• Where did students have the most opportunity to engage?
• Which instructional strategies showed up in practice?
• What could I adjust the next time I teach this lesson?

This kind of feedback is especially powerful because it is timely, contextual, and immediately useful. Teachers can reflect while the lesson is still fresh and make practical adjustments for tomorrow.

What we learned from our pilot

Before releasing this feature publicly, we piloted it with ESUCC, a state education agency in Nebraska. Teachers uploaded a series of lesson plans, used M2 while teaching, received live tips, and then reviewed feedback that was detailed and aligned to their plans.

The pilot confirmed what we hoped: when M2 understands the lesson plan, the feedback feels more relevant to the teacher’s actual classroom experience.

One teacher using the Wit & Wisdom curriculum shared that “the feedback was very specific to this particular lesson” and that “most of my feedback was very specific to a certain question or task that I asked.”

That is exactly the kind of support we want M2 to provide: not abstract feedback, but feedback that helps teachers see their own lesson more clearly.

Why this is exceptional

AI in education is only useful when it understands context. M2 is designed to connect what was planned, what happened during instruction, and what a teacher can do next. Lesson plan upload makes that connection stronger.

Teachers can enter the classroom with a plan, receive support while teaching, and leave with feedback that helps them reflect, adjust, and grow. For administrators and coaches, this creates a clearer bridge between curriculum implementation and classroom practice.

That is what makes this an exceptional addition to M2: it brings planning, teaching, feedback, and improvement into one more coherent experience.

The post Bring your lesson plan into M2: More coherent support before, during, and after teaching appeared first on Swivl.

If you’ve been through the National Board process, or supported someone who has, you know the scope of the commitment. It’s an exhausting, rewarding marathon of recording lessons, rewatching them, and trying to self-assess against demanding rubric criteria.

For many teachers, the hardest part is the isolation. With Swivl’s M2, candidates gain a comprehensive preparation system that supports them through the messy middle of the process, building the awareness and skills needed to produce a high-scoring portfolio.

Phase 1: Building awareness through structured feedback

National Board commentaries ask candidates to narrate specific instructional decisions. Reconstructing these decisions from memory is a challenge, especially if a teacher is reflecting days or weeks after a lesson.

M2’s My Feedback breaks practice sessions into segments (opener, direct instruction, guided practice, etc.), providing a well-organized map of the lesson. This helps candidates pinpoint exactly when instructional goals were met or missed.

By providing far more structured feedback than teachers would typically get during a school year, M2 helps National Board candidates build their skill of metacognitive reflection, which transfers directly to their submission work.

Phase 2: Targeted self-assessment

The National Board is not only about good teaching; it is also about meeting specific, rigorous standards.

Using the Rubric Builder, candidates or coaches can create criteria modeled directly on National Board standards, such as differentiation or student engagement. Then, a teacher simply turns on M2 at the beginning of any class, and they’ll receive a detailed feedback report based on the National Board rubric criteria.

By self-assessing their practice and making instructional adjustments, candidates can see their scores trend upward over time. When candidates are then ready to record and submit their final submissions, they come to that experience with a sense of confidence built over weeks or months of deliberate practice.

Phase 3: Internalizing the reflective cycle

Reflective practice is the through-line of the entire National Board process. Chat with M2 offers a coaching-style conversation after practice sessions, drawing on transcripts to help candidates talk through their reasoning: Why did I shift strategies there? How would I justify this to an assessor?

M2 provides a unique opportunity for a teacher to have a detailed, evidence-based conversation with a voice that has knowledge of the lesson that was just taught. Furthermore, a teacher can have this experience every day, or even multiple times a day, when they are working with M2.

In early practice sessions, Live Tips act as a digital coach, offering classroom management tips, check-for-understanding suggestions and discussion questions in real-time.

Mid-lesson, a teacher can request a tip from M2, which is tailored to the precise context of the class being taught. The teacher can then implement this adjustment to their lesson and monitor the results. This builds the “muscle memory” for the reflect-adjust-teach loop required by the Board. By the time a candidate is ready for their final submission, these high-impact moves have become second nature.

Phase 4: The final submission

When a candidate has built the confidence and metacognitive skills to meet the standards, the M2 hardware transitions from a coach to a professional recording suite.

For final submissions, candidates can turn off the AI features and use the M2’s high-fidelity audio and tracking to capture the continuous, unedited video required by the Board. Because they have practiced with the same hardware for months, the camera becomes “invisible” to both the teacher and the students, resulting in a more authentic, natural classroom environment.

Preparation that compounds

The National Board process is designed to identify teachers who already practice at a high level. M2 doesn’t do the work for the candidate; it helps the candidate make their practice visible. By the time teachers hit “submit,” they aren’t just hoping for a high score. Instead, they have the data and the habits to know they’ve earned it.

The post Beyond the video submission: Why M2 is a system for National Board success appeared first on Swivl.

From the beginning, we made it clear that teachers knew M2 would not be about evaluating instruction. It’s for ‘making me a better teacher and seeing those results in my student outcomes’.

Brian Teague | Bowie Elementary, Lamar CISD

USE CASE

Instructional coaching and student outcomes

Katherine Lange

Principal

School/District:

Bowie Elementary, Lamar CISD

Grade Level

K–5 Elementary

One new principal, four instructional coaches, and a lot of new teachers

Katherine Lange had never led an elementary school before she walked through the doors of Bowie Elementary in January 2025, and this particular elementary school had been struggling. 

Tier I Instructional practices were inconsistent, educator turnover had been significant, and the year ahead was shaping up to require more than just a few steady hands at the wheel. Katherine has four instructional coaches on her team, but she was also facing how she would effectively support a brand-new crop of teachers.

“Last year was all about triage. We came out of it earning a B, which we’re all proud to accomplish,” she beamed. “Coming into this year, it needed to be less about triage and more about making targeted instructional shifts. Yes, we wanted to sustain the grade, sustain the growth, but we needed to really expand what teachers can do with the kids in the classroom.”

Students first

One of Katherine’s primary goals was to prepare students for extended constructed responses on Texas’s STAAR exam. She was seeing a trend emerge that students could find textual evidence, but they somehow circumvented answering the question explicitly, and on the STAAR, that could give a student an automatic zero on those questions. To help students improve, M2 was given a spot in many classrooms’ studio rotations (aka: small group stations). Set up this way, teachers could define an ECR as a quick station prompt within M2, then students could orate their written responses and M2 would offer immediate feedback to the students directly.

And when studios were over, teachers could later review a full summary of what had happened across every student interaction, using the data to monitor and adjust instruction and ultimately: improve students’ ECRs on STAAR.

For teachers, M2 narrowed the ever-widening gap in academic feedback. Erin Lundberg, a fifth-grade teacher, had tried lots of strategies to offer daily feedback: exit tickets, ghost rotations, checklists. But what it boiled down to was: she just didn’t have enough time each day to check in on each student.

“When I’m fully engaged with the studio I’m running, I’m unable to listen in on the conversations happening at other studios, so I can’t quickly readjust a student’s thinking or address their misconceptions in the moment. That’s where M2 comes in! He is my teacher partner, and the students check in with him when I can’t.”

For Erin’s emergent bilingual (EB) students, M2 helps her keep a fresh list of never-ending sentence stems to increase their language acquisition. For everyone else, M2 drives student inquiry and increases Erin’s higher-order questions.

Carmen Chavero, who teaches fourth-grade dual language math and science, faced similar constraints. Students working independently while she ran small groups had limited access to daily feedback. As Carmen puts it “M2 has been like having a second teacher. I love how it provides meaningful feedback for every student and is constantly offering them new and creative ways to approach and solve math problems.”

Privacy paves the way

Before M2 became employed for a coaching cycle, Katherine established one non-negotiable rule: all coaching data would remain strictly between the instructional coaches and the teachers they worked with – never used for evaluation, never surfaced to administration (and by default, M2 data is stored in a teacher’s account, not an administrator’s).

That decision, made publicly and kept consistently, fundamentally altered the way teachers perceived the tool.

Brian uploaded LCISD’s instructional look-fors and the T-TESS domains and dimensions framework directly into M2’s customizable platform, so every piece of feedback M2 generated was grounded in the language teachers already used and the standards they were already being measured against. What results is feedback that feels comforting and familiar. 

And that’s when teachers really begin to believe in the tool and advocate for it.

“As the instructional leadership team who purchased M2, of course we can say: ‘yes! this is a great tool.’ But when our teachers talk to each other and say, ‘no, really, it is good, you should use it,’ that always carries more weight,” muses Katherine.

M2 as a coaching partner

One of the more unexpected uses of M2 at Bowie wasn’t in a classroom at all. It was in a coaching conversation.

Brian began bringing M2 directly into his debrief sessions with teachers. He’d start the coaching conversation, work through the normal identifying questions, and then, when a question came up that needed a second perspective, he’d turn to M2. He’d give his input, then ask M2 what it thought. Or he’d let M2 respond first and add to it. Either way, the result was the same: a third voice in the room, grounded in the district’s own framework, providing objective reinforcement to the conversation.

For Brian, who was himself new to the coaching role, this mattered in ways that went beyond any individual teacher conversation. Hearing how M2 framed feedback on engagement, pacing, or rigor helped him develop his own coaching language.

M2 extended into planning time, too. Brian brought the robot into collaborative meetings to generate ideas for differentiation. Certain teachers who might have felt guarded in a one-on-one coaching session found the dynamic felt extremely lighter when these conversations included a third, objective presence.

Greatest risk, even greater rewards

One new teacher came into his first coaching cycle at Bowie hesitant. He was nervous about hearing or seeing himself (video is always optional in M2), uncertain about what he’d uncover after reviewing each lesson.

By the end of the year, he was asking to keep M2 for extended periods of time, to cover entire units. He wanted to see where his engagement dipped, and why. He wanted to connect what M2 was showing him to his own lesson plans, find the gaps, and fix them.

His student data reflected the shift. 

“He was able to go back and look at his trends and find places where he could grow,” Katherine explains. “It really helped him be more prescriptive in the way that he planned.”

He wasn’t the only one. Even master teachers were experiencing their own magical moments with M2. When you walk into a strong teacher’s room, you tend to see strengths but what you can miss are edge cases where strong – but common – instructional moves aren’t quite nailing it, where you need to really try an out-of-the-box strategy.  

“It provided insights for a master teacher that instructional coaches and administrators sometimes even overlook,” Katherine says,

What do Title I funds and screen time limits have in common?

Because M2 addressed objectives embedded directly in Bowie’s campus improvement plan – recruitment and retention of high-quality teachers among them – Katherine was able to fund every M2 purchase through Title I dollars. The investment had a clear, documented justification.

There was one more constraint she hadn’t anticipated finding a solution to: the district had recently implemented screen time limits for early elementary students. This is where Swivl and the district had the same intentions – and stars aligned. With M2, students and teachers are able to interact with the robot conversationally, and its remote microphones prevent anyone from becoming glued to its screen. M2 even intentionally designs its classroom activities and group experiences to promote conversations and emphasize paper-based work time, so it ended up being a natural fit that did not compete with the district’s mandate. 

Bowie’s ‘B’ was for believing

Before M2, some of Bowie’s newest teachers were anxious. About being recorded. About AI in class. About what the data might reveal.

After a few short months, many of those same teachers were requesting M2 on their own, asking to keep it for full units, and encouraging their colleagues to try it. The anxiety hadn’t disappeared because Brian and Katherine reassured them everything was safe and private (though that helped, of course). It disappeared because they saw visible changes in their practice, session by session, trend by trend, unit by unit – and they wanted more of it.

With student outcomes being the primary motivation for Bowie’s teachers, they’ve come to rely on M2 to be their classroom “wing-robot”. Carmen Chavero envisions that next year, M2 will be a permanent workstation: a standing second small group, always available, always ready for the next student who needs a different explanation or a patient ear. Erin Lundberg is already using what M2 shows her to adjust pacing, deepen questioning, and ensure no student falls through the cracks of a busy rotation. And she’ll come back even stronger in the Fall.

This is exactly what Katherine had in mind when she arrived at Bowie mid-year, inheriting a campus in triage. Their mantra has been “keep your eyes on the A and it’ll be ours in May!” They continue to soar because they built a culture where everyone on campus has a chance to grow, become more accountable, and thrive as learners. No small order, but it just so happens that M2 helps them do that.  

The post #OneLittleSpark: How Bowie Elementary kept its flame aglow with M2 appeared first on Swivl.

What does a coaching model look like when you have five middle schools, a limited budget, high teacher turnover, and academic coaches and school administrators who can’t be everywhere at once? 

For Dr. Loralee Price, Instructional Technology and Data Integration Supervisor for Hawkins County School District in rural East Tennessee, finding the answer to this question was the district’s daily struggle.

Like many districts, Hawkins County has dedicated academic coaches and administrators who care deeply about teacher growth. But care and capacity are two concepts that sometimes fail to become compatible. Coaches split across schools could observe a teacher once or twice a month. Administrators walk through classrooms weekly. But by the time a meaningful debrief occurs, that lesson might be a distant memory. 

The district needed something that could close that gap: feedback that was immediate, specific, and available in every classroom, not just those a coach or administrator was able to visit.

The other challenge is one that many districts in the region are experiencing: constant turnover. A wave of newer educators has flocked to Hawkins County’s middle schools in recent years, many with limited classroom experience. For those teachers, the traditional model of professional development: periodic observations, scheduled debriefs, formal evaluations, rinse/repeat, can feel like too little, arriving too late.

Budget constraints compounded the issue. Hawkins County is a small, rural district where, as Dr. Price puts it, “funding is scarce.” Expanding professional development in any meaningful way meant finding a solution that could scale across five schools without requiring a proportional increase in cost or staffing.

When the district learned about M2, they scraped together what limited budget they could to invest in hardware. And upon learning about Swivl’s AI Innovation Grant in 2025, they applied for and won the remainder that they needed to realize their first full pilot year. With the assistance of grant funds, they were able to ensure that every middle school, every core and special education teacher, every coach and administrator, in the program from day one could be supported by M2. Here’s what they accomplished in that first year.

Noticing the unnoticed

The first thing teachers noticed was that M2 was paying attention to things that no one else was able to spend enough time to notice themselves. 

Sydney Allen, from Church Hill Middle School, moved to the far side of her classroom mid-lesson to check in quietly with a student who was struggling with material. It was the kind of moment that is often missed: too small for a formal observation, too easy to forget by the time a coach arrives. M2 picked up the conversation, provided some concrete next steps for the teacher in her post-session feedback, and bam! A situation that could have led to a student falling through the cracks was rescued by M2’s reminder. 

Wesley Thomas, of Rogersville Middle School, discovered something he hadn’t known about himself. “After reflecting on a lesson with M2 advised I use a timer for certain portions of class. I don’t think I would have realized this on my own.” A pacing problem, invisible to him, made visible by a single reflection with M2.

Shawn Swickheimer, in his second year at Church Hill Middle, found a different kind of value. By comparing lesson summaries and takeaways across class periods, he could see which class hadn’t fully grasped the material and adjust before teaching it again the next day.

Elizabeth Saxena at Church Hill Middle used M2 during a formal evaluation. M2’s scoring aligned closely with the formal evaluator’s score. For teachers like her who can feel anxious about formal observations, that reassurance can make all the difference to preparing for the real thing.

M2 in the classroom: Real examples

The impact wasn’t limited to teacher growth. In Jeffrey Klepper’s ELA classroom at Surgoinsville Middle School, students used Ask M2 to strengthen how they cited evidence from texts — a literacy skill that Klepper was actively building, with M2 reinforcing it in the moment students needed it. 

In Hannah Price’s science class at Church Hill Middle, students weren’t waiting to be called on. They were asking M2 questions on their own, driven by curiosity that the lesson had sparked. 

At Clinch School, Jessica Drinnon’s ELA students worked through small group discussions with M2 providing support alongside them.

Skepticism turned into a superpower

Before any of these gains could be realized, teachers had to trust the tool. That’s rarely a given, especially with technology that assesses instruction.

Heidi Hesoun at Surgoinsville Middle School was one such teacher whose trust M2 needed to earn. “When our district team introduced M2, I initially had a lot of reservations, but I gave it a chance, and I immediately found value in its feedback and instructional ideas. In fact, I look forward to bringing it to my class because I know it will give me inspiration that is actually relevant and helpful for what I am teaching that day. The best part is the instant feedback I can get after every lesson. I don’t have to wait to schedule a meeting for someone to observe me, it’s my own little ‘coach.’”

Heidi’s principal, Karen Bear, also noted the initial hesitation around the school. “Teachers in my building were initially skeptical due to various concerns about who might see the lesson; however, through consistent use and an understanding that all of their lesson data would remain private, they have come to embrace it as an everyday tool that enhances their instructional practice.”

Many teachers felt like Heidi at first, so during one of the district’s onsite trainings, Swivl’s professional development team ran an exercise called “What M2 is NOT Saying.” The goal: help teachers understand how to read M2’s feedback without getting stuck on the details and inevitably spiraling. For example, feedback from M2 that surfaces a pattern in someone’s pacing or questioning isn’t a verdict. It’s simply a data point to grow from. And the point was driven home: your lesson – your data – your privacy. You choose what and when to share. The exercise reframed what M2 observed from something evaluative into something useful.By the end of this crucial session, every teacher had created a personal goal tied to their own data and a concrete coaching plan for acting on it. The room that had walked in with questions walked out with direction.

What’s next for Hawkins

Dr. Price can attest that “by piloting this technology in our middle schools, we’ve moved from a model of occasional observations to a culture of continuous reflection and improvement.” Her vision for M2 extends beyond just a middle school pilot. 

What she has found most encouraging is how the ‘privacy-first’ design of M2 shifted teacher mindset from evaluation to empowerment. Dr. Price notes “our novice teachers, in particular, are using this non-biased feedback to master instructional skills in a low-stakes environment. As we look toward scaling—even amidst budget constraints—we are prioritizing our early-career educators. We aren’t just giving them a tool; we are giving them the autonomy to own their professional growth.”

For any district, rural or otherwise, where budget constraints are constant and capacity is finite, the value of a tool that’s present in every classroom cannot be understated. Catching the quiet conversation in the corner, surfacing a pacing problem the teacher didn’t know they had, helping someone see what’s working and what isn’t before another week passes. Without a constant companion to help teachers grow, these issues pile up, feedback falls through the cracks, and students ultimately suffer.

Teachers can’t see everything happening in their classrooms. Coaches can’t be everywhere all the time. Admins can’t be clairvoyant every day. But M2 can.

The post From grant to growth: How Hawkins County brought AI innovation to every middle school classroom appeared first on Swivl.

Picture a student teacher the morning of their first formal observation. They’ve prepared for days. The formal observations feels high stakes because in a sense it is: observations are rare. Not because anyone has failed them, but because the math of field placement supervision makes frequent feedback hard to deliver.

A university supervisor managing a full roster of student teachers across a dozen school sites, balancing their own teaching load, can often realistically visit each candidate two to three times a semester. 

Between those visits, student teachers do get support, but it comes with real limits:

• Mentor teachers are present and often helpful, but their feedback reflects a teacher working within one building’s context, not necessarily the frameworks a preparation program is building toward
• Cohort meetings and seminar discussions give candidates space to process, but the feedback is secondhand. No one else was in the room.

What if that gap didn’t have to define the experience? How could student teachers receive specific, framework-aligned feedback after every session? Could supervisors arrive at observations with a semester’s worth of data, rather than reconstructing a narrative from two visits?

Fortunately, there’s a way to augment the traditional model to make this all possible.

The problem with episodic observation

When a student teacher knows they’ll be formally observed once or twice this semester, that session carries weight. They prepare differently, perform. receive feedback that reflects a curated version of their practice, not necessarily the daily reality of their work.

Practically speaking, issues that surface in week three of a placement may not get identified until a supervisor visits weeks later. By then, the teacher candidate may calcify certain patterns that are causing persistent issues in classroom management, instructional planning, or elsewhere. All the while, a small course correction could have prevented the larger issue from emerging. The model also asks supervisors to draw broad conclusions from a narrow slice of evidence: two observations, however carefully conducted, are snapshots, not a portrait.

While issues with infrequent feedback are logically clear, the benefits of frequent feedback are research-backed.

• John Hattie’s Visible Learning synthesis puts feedback at an effect size of 0.70, among the highest-impact interventions identified across decades of educational research. 
• Wilcoxen and Lemke reached a similar finding: ongoing, process-focused feedback raises performance in ways that summative evaluation simply cannot.
• Schaefer and Clandinin found that regular observations with constructive feedback were highly valued by beginning teachers
• Garcia and Weiss reported that the highest-performing education systems in the world provide feedback as part of the regular experience of teaching.

A key theme in this body of research is regular and ongoing. Feedback works when it’s frequent, specific, and connected to real practice.

For student teachers, the implication is direct: a placement evaluated through two observations is, by design, a low-feedback environment, which is less conducive to growth.

What three to four months of M2 looks like instead

M2 gives student teachers something the traditional model doesn’t: a feedback loop that runs alongside their practice rather than interrupting it once or twice a semester.

When student teachers use M2 weekly, they receive AI-powered feedback after every session on:

• Questioning techniques
• Participation patterns
• Instructional pacing
• Classroom dynamics

They start identifying their own tendencies before a supervisor ever needs to name them. Early in a placement, when there’s the most room for improvement, that kind of immediate, low-stakes reflection has an outsized impact.

A typical M2 feedback session might surface something like:

*”Students responded well to your opening question, but whole-class participation dropped during the independent work phase. Consider building in a structured think-pair-share before returning to whole-group discussion.”**

That’s the kind of specific, actionable language student teachers can act on immediately. It also maps directly onto the competencies supervisors are already evaluating. M2 feedback aligns to standard teacher preparation frameworks like Danielson and Marzano, connecting AI insights to the rubrics programs are already using.

M2 doesn’t replace formal observations or human judgment. What it gives supervisors is a richer evidence base: a complete arc of real sessions, real classrooms, real decisions, not two curated clips. When coaching conversations are grounded in session-by-session data rather than one visit from three weeks ago, they become more specific, more productive, and more useful to the student teacher sitting across the table. Dr. Natalie Bolton’s experience shows what this looks like in practice.

The habit that carries forward

The question a pre-service program has to answer is clear: do our teachers leave with the skills needed to teach today, and the mindset to get better tomorrow?

Teaching rewards people who treat their practice as something worth examining. The teachers who grow most consistently over a career are those who stayed genuinely curious about their own work. That disposition is easiest to form at the very beginning, before other habits are established. 

The structure of a field placement helps to shape it. When student teachers receive feedback weekly, they internalize a simple but durable lesson: looking closely at their practice is a normal part of the job. When formal observations are the only occasion for feedback, they internalize that evaluation is rare, high-stakes, and something to prepare for rather than learn from. Whichever lesson they carry out of a placement, they’re more likely to carry forward.

A semester of weekly M2 sessions is a semester spent building that habit. The argument for continuous feedback goes beyond this semester. It’s a foundation for everything that comes after.

The post Why 1-2 observations aren’t enough: The case for continuous feedback in pre-service teacher training appeared first on Swivl.

At Swivl, we take this seriously. This post explains what we’ve done to ensure M2 scores like a trained human evaluator, and how we continue to monitor and refine that accuracy over time.

M2’s Custom Rubrics

M2 provides organizations a way to define the instructional criteria that matter most to their team. Through M2’s custom rubrics feature, organizational leaders can specify performance indicators and scoring criteria that reflect their values and expectations. 

After each lesson, M2 evaluates the teaching activity against those rubrics and returns a score on a 1–4 scale, along with feedback explaining the score, highlighting strengths, and identifying areas for growth.

The promise of this feature is powerful: frequent, consistent, criteria-aligned feedback at scale, without walk-throughs or high-pressure observations. But that promise only holds if the scores are accurate.

The problem we set out to solve

Early in M2’s development, we identified a scoring distribution challenge: the system was returning scores that clustered in the middle of the scale, producing mostly 2s and 3s regardless of actual instructional quality. This is a common challenge when building AI scoring systems. The model hedges rather than discriminates, and the result is feedback that feels generic and uninformative.

We needed M2 to produce a score distribution that looked like what trained evaluators produce, which is an appropriate 1–4 spread that reflects genuine differences in instructional performance.

How we calibrated M2 against expert evaluators

To address this, we built an evaluation dataset using classroom observation videos that had already been scored by highly trained human evaluators. These were not casual reviewers, but evaluators with deep familiarity with professional teaching rubrics and established inter-rater reliability.

We then ran M2 against those same videos and compared the output.

The gap between the two distributions was clear. Our human evaluators produced the kind of spread you’d expect from a well-calibrated rubric: a meaningful range of 1s, 2s, 3s, and 4s. M2’s initial output clustered near the middle.

From there, we went through an iterative calibration process: adjusting how M2 analyzes transcripts, restructuring the way rubric criteria are applied, and refining the scoring methodology at each step. Each iteration was tested against the same evaluation dataset and compared to the human baseline.

One key improvement came from how we structured M2’s rubric evaluation process. Rather than evaluating a lesson holistically, we introduced a more structured, criterion-by-criterion approach. Essentially, we gave M2 a more disciplined framework for applying each rubric dimension, similar to how a trained observer would work through an evaluation instrument item by item.

We also differentiated M2’s feedback language by score level. A score of 1 now more clearly focuses on what was missing and what the teacher can do differently. A score of 4 now emphasizes what went well and why it was effective. This mirrors how skilled coaches communicate; the message you deliver to a struggling teacher is structurally different from the message you deliver to a strong one.

After calibration, M2’s score distributions align closely with those of our human evaluators. The gap narrowed substantially, and M2 now produces the range and differentiation that makes feedback actionable.

What the remaining differences tell us

The goal of this calibration was less about achieving a perfect match, and more about understanding where differences exist and why. There are cases where M2 and human evaluators diverge, and those cases are instructive.

Most remaining differences occur at the boundaries of rubric descriptors: situations where a teaching performance sits between a 2 and a 3, for example, and reasonable evaluators could score it either way. This is not a failure of the AI. Instead, it reflects the same ambiguity that human evaluators navigate. 

In fact, inter-rater reliability among trained human evaluators on rubric items like these is rarely 100%, and M2’s agreement rate with our evaluators compares favorably to the agreement rate between two independent human scorers on the same material. 

For now, one unique capability of human evaluators is eyesight. M2 currently relies on class transcripts for the analysis that leads to scores and feedback. However, we are investigating the safe, privacy-centric approach to bringing visual evaluation capabilities to M2 in upcoming releases.

Understanding boundary cases and the work process of AI vs. humans helps us continue to define clear paths for improvement.

Ongoing monitoring and refinement

Calibration is not a one-time event. As M2 is deployed across diverse classrooms, grade levels, and instructional contexts, we continue to monitor scoring accuracy and precision.

We define accuracy as how closely M2’s scores align with a trained human evaluator. We define precision as how consistent M2’s scoring is across similar lessons. In other words, whether it gives comparable scores when evaluating comparable teaching. Both accuracy and precision matter.

Our team runs ongoing comparisons between M2 output and human-evaluated samples, identifying drift and opportunities to improve. When we find systematic gaps, we refine our approach and re-validate. This continuous loop is what allows M2’s scoring to remain trustworthy over time, not just at the point of initial release.

The result: consistent, trustworthy feedback

When an organization deploys M2 with custom rubrics, they are not getting a black-box AI generating arbitrary scores. They are getting a scoring system that has been explicitly calibrated to match how trained evaluators apply rubric criteria — and that is continuously monitored to stay that way.

This is what makes M2’s frequent feedback valuable. Not just that teachers receive more of it, but that the feedback they receive is grounded in the same framework a skilled evaluator would apply. The score of 3 a teacher receives on Monday should mean the same thing as the score of 3 they receive two weeks later, because M2 applies its criteria consistently.

When teachers and leaders can trust the scores, they can use them. Coaching conversations become more specific. Professional development becomes more targeted. And leaders gain a reliable, ongoing view of where their organization is performing and where it needs to grow — without the cost and inconsistency of traditional evaluation systems.

M2’s AI rubric scoring is built to earn that trust. And we’ve done the work to prove it.

The post How M2’s AI rubric scoring aligns with expert human evaluators appeared first on Swivl.