← All education signals
Education

Pace of One

The tutor built around a single learner has arrived ahead of its proof: one randomised trial finds its gains resemble practice without AI, and most teachers get no guidance.

What is it?

Tutoring software built around one student’s pace is in classrooms, but a two-year randomised trial finds its maths gains resemble practice without AI. Most teachers get no guidance on using it. Whether the tutor earns its place, or stays an unopened tab, is the pivot.

Why now?

Fixed pacing was always a proxy for a cost nobody could avoid: one teacher, one lesson, one class average. Models cheap enough to tutor one student at a time challenge that, yet a two-year randomised trial finds no clear edge over practice without AI. The bottleneck moves from capability to whether students use it and teachers know how.

Who’s building it?

Edtech platforms building AI tutors for one learner at a time, now facing independent randomised trials of their gains.

Khan AcademyEdtech vendorsSchool districtsTeachersResearchers
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — AI-tutor platforms whose gains are tested in randomised trials.
Buyers — school districts and teachers, mostly without guidance on AI tutoring.
Gatekeepers — accreditors who still define learning by time, not skill.
What’s at stake
Credentials and enrolment periods built on a shared clock lose meaning once pacing isn’t fixed — the calendar is downstream.
A randomised trial finds AI-tutor gains resemble plain practice.
A tutor helps only on the days students open it.
What’s unsettled
Whether accreditation adapts to skill pacing, or keeps the fixed clock.
Whether procurement rewards proven trial evidence, or still follows vendor narrative.
Whether teachers get guidance on AI tutoring, or keep deciding alone.
02 · How the demand matureda progression, not a timeline
Pilot
Schools trial AI tutors with students, as in a two-year randomised study across 18 Tennessee middle schools.
Evidence
Trial gains with the AI tutor resemble those from practice without AI.
Procurement
Districts move spend into standing curriculum lines only once trials show gains beyond plain practice.
Standard
Skill-based pacing becomes the expected default, unsettling credentials built on fixed time.
03 · Where the market goesthe forks, not a prediction
→Skill pacing or fixed clock — decided by whether accreditors adapt credentials to match, or hold the calendar in place.
→Evidence-led procurement could fully replace pitch-led buying across the whole edtech category, eventually, everywhere.
→Students take up the tutor as a daily habit, or leave it unopened on most practice days.
→The range of ability inside one grade keeps forcing the pacing question back onto districts.
The premise question underneath: does a skill-based unit actually serve the learner, or just make procurement easier to justify on paper?
04 · The Bespoke Default deskadjacent markets, moving now
The Always-On Tutor
A pilot lifts scores by years.
AI tutoring lifted scores 1.5-2 school years in 6 weeks.
World Bank
The Tutor Nobody Opens
Access alone rarely guarantees any real use.
Only 15% of students with access actually use it.
EdTech Innovation Hub
Generative Medicine
The same bespoke logic reaches drug design.
AI-designed drug rentosertib entered Phase 3 trials.
ClinicalTrials.gov
05 · The Wire
01
Randomised trial: AI tutor gains resemble Khan Academy practice without AI
NBER (working paper) ↗
02
69% of US teachers get no guidance on AI for tutoring
Gallup ↗
The panel on this signal
They agreed on
–Independent randomised trials are the fair test of an AI tutor’s claims, and one finds no edge over plain practice.
–Pacing by the individual is the promise, not yet the measured result.
–Teachers are left to fit AI tutoring into lessons with little or no guidance.
The learning scientist disagreed on
–Institutional accreditation structures remain built around fixed time-based units; whether skill-paced tools ever translate into an accreditation system that measures competency, rather than being absorbed within an unchanged seat-time framework, is unresolved.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.
✦Signal changelogre-read as evidence changed
2025
Khan Academy’s tutor shows a measurable accuracy lift once it uses a student’s learning history, and personalised learning is moving into procurement.
→
2026
An independent randomised trial found gains from Khan Academy with its AI tutor resembled gains from the same practice without AI assistance. NBER (working paper) ↗