Logan & Friends
| 2024
Delivered near real-time instructional coaching and cut feedback navigation time 40% without letting AI override human judgment
Impact
Efficiency rate
100%
Coaching Feedback Moved From Delayed → Near Real-Time
Eliminated multi-day feedback delays, enabling teachers to reflect while lessons were still fresh.
User satisfaction
90%
90% educator satisfaction with clarity and speed
Trust increased when AI insights were contextual, optional, and reviewed by a human coach.
Navigation time
-40%
Reduced feedback navigation time by 40% (6 min → 3.5 min)
Teachers spent less time searching and more time reflecting and improving potentially.
Context
Logan & Friends is an education company that partners with schools to design equity-focused, real-world learning experiences. A core part of their work involves instructional coaching, observing classroom teaching and providing feedback to help educators grow.
Instructional coaching works in theory.
In practice, it doesn’t scale well.
Feedback often arrives days or weeks after a lesson, long after its impact has faded.
The client’s vision was ambitious:
to explore whether AI could reduce feedback delays by analyzing voice-recorded teaching sessions, generating standards-aligned insights, and supporting human coaches in delivering faster, more actionable guidance.
However, the scope, technical feasibility, and trust boundaries of using AI in classrooms were unclear.
TL;DR
Instructional coaching doesn’t fail because feedback is bad.
It fails because it arrives too late, without context, and without control.
As one of the product designers on the team, I helped design an AI-assisted coaching system that prioritized timeliness, clarity, and human judgment over automation.
By anchoring AI insights to real classroom moments and tightly constraining when and how AI intervened, we reduced navigation friction, eliminated coaching delays, and preserved trust.
Role
Product Designer
Team
4 Product Designers · Client Designer (PhD) · Subject Matter Experts
Domain
EdTech · Human-Centered AI
Timeline
30 Weeks
Problem
The surface problem was delayed feedback.
The deeper problem was misalignment between how coaching tools are designed and how teachers actually reflect and improve.
What wasn’t working
Feedback often arrived 72+ hours after observation
Guidance felt generic and disconnected from real classroom moments
Teachers struggled to review long, dense feedback within busy schedules
Existing tools added cognitive load instead of reducing it
At the same time, introducing AI added new risks:
Teachers feared being judged or evaluated by algorithms
Recording classrooms raised privacy and consent concerns
Automation risked removing the empathy central to good coaching
The challenge wasn’t whether AI could generate feedback.
It was whether it could do so without breaking trust.
Research
To ground the product in real educator needs, I led and synthesized research across multiple methods.
Research methods
5 in-depth teacher interviews
12 educator survey responses
7 competitor analyses
Review of 20+ research papers on instructional coaching and classroom observation
Multiple rounds of usability testing on early concepts
Key research insights
1
Timeliness matters more than volume
Teachers consistently said feedback loses value when it arrives days later, even if it’s detailed.
2
Context builds
trust
Educators were far more receptive to feedback when it was tied to specific classroom moments they could recall.
3
AI acceptance is conditional
Most teachers were open to AI-assisted feedback only if it didn’t feel judgmental, intrusive, or opaque.
4
Navigation friction kills adoption
Usability testing showed teachers spent excessive time finding relevant feedback, increasing frustration and disengagement.
Empathy and affinity mapping reinforced a clear pattern:
Teachers wanted bite-sized, contextual guidance that respected their time and professional judgment.
The Guiding Principle
AI should surface insight, not deliver judgment.
This principle guided every design decision.
AI’s role: listen, identify patterns, surface moments worth attention
Human coaches’ role: contextualize, interpret, and guide growth
Early Direction: Insight Summaries
I designed the feedback screen to give teachers a clear snapshot of summarized session insights into categorized cards.
Feedback is grouped into clear categories with color-coded tags (1) that highlight framework patterns, with simple, action-focused notes (2), making it feel like guidance rather than evaluation. and a “View related resources” link (3) provides deeper context and help without cluttering the view, keeping feedback approachable and scannable.
What worked
Clear structure
Fast overview
What failed
Teachers couldn’t tell why feedback appeared
No connection to specific classroom moments
Feedback felt generic, the same frustration teachers already had
Decision 1:
Anchor Feedback to Real Classroom Moments
Based on research and testing, I redesigned feedback to be
time-based and contextual.
Switched to a vertical, chronological layout
Added timestamps tied directly to the lesson
Renamed “Session Insights” to "Action Items"
That way, teachers can instantly see what each note refers to, why it was flagged and implement those actions in practice.
Before
After
Anchoring feedback to real moments made AI feel observational, not evaluative.
This was the first major trust unlock.
Decision 2
Despite improved context, testing showed teachers spent over 6+ minutes navigating feedback.
So I redesigned the feedback experience to be discoverable, scannable first, deep second.
Added an audio player with timeline markers
Allowed quick jumps to relevant moments
Why it mattered
Teachers are time-poor. If value isn’t easy to find, intelligence doesn’t matter.
Decision 3
Make AI Optional, Contextual, and User-Invoked
Early research showed teachers were open to AI, but only if it didn’t interrupt or judge them.
What changed
AI chat was added as an on-demand assistant
It responded only to questions about selected feedback
It never surfaced unsolicited evaluations or scoring
Why it mattered
AI became a support tool, not an authority. Teachers stayed in control of when, how, and how much they engaged with AI.
This preserved the human core of coaching while still gaining speed.
The Final Experience
The final prototype combined:
Timestamped feedback tied to real classroom moments
Audio playback with visual markers
Clear, scannable feedback cards
Optional AI support for clarification and exploration
Human coach review before delivery
The system felt less like an evaluation tool and more like a coaching partner.
Validation & Testing
We validated clarity, efficiency, and trust through:
Expert reviews with instructional coaches
Think-aloud sessions with experienced teachers, novice teachers, and coaches
Cognitive walkthroughs and design critiques
Post-test satisfaction questionnaires
Key findings and design responses
Issue:
The Audio player remained fixed on top that reduced the usable scroll area, constrained scrolling made longer feedback hard to review.
Before
Fixed:
I reworked the layout to prioritize reading and flow by resizing the existnig containers, expanding the primary scroll area to allow continuous reading and made the audio player sticky.
After
Execution
I led end-to-end design from concept to validation:
Defined information architecture and feedback flows
Designed low- to mid/high-fidelity prototypes
Ran multiple rounds of usability testing
Iterated based on observed behavior, not assumptions
Collaborated closely with designers, a PhD client, and subject matter experts to ensure pedagogical and technical feasibility
Design decisions were continuously tested and refined to balance clarity, trust, and speed.
Impact
Usability testing and prototype validation showed measurable improvements:
40% reduction in navigation time (6 min → 3.5 min)
Near real-time feedback, eliminating multi-day coaching delays
90% of teachers reported satisfaction with clarity and speed
Increased confidence in AI-assisted feedback when paired with human review
Most importantly, teachers described the experience as supportive, not judgmental.
AI Design Trade-offs
We intentionally traded:
Flexibility for trust
Automation for clarity
Intelligence for restraint
AI worked best when it:
stayed close to real moments
avoided abstract scoring
respected human authority
My Learnings
This project reinforced that designing AI systems is as much about restraint as capability.
1
Trust is earned through predictability, not novelty
2
AI adoption depends on timing, context, and control
3
Human judgment must remain visible in high-stakes domains
4
Iteration is how ambiguity becomes clarity
What’s Next
1
Expand coaching insights across multiple sessions
2
Strengthen privacy and consent controls
3
Validate long-term impact on teacher growth
4
Explore real classroom pilots