The argument writes itself. A UK independent day school costs £15,000–£25,000 a year; a boarding school £35,000–£55,000. Against that, an AI tutor — whether a £14 one-time top-up for a month of practice or coach-led PRO tuition at £360/term — looks like rounding error.

But "looks small" isn't an answer. The honest question is: what does that spend actually buy you, and how does that compare to the alternatives? This article is the working.

The comparison set

To answer "is it worth it", we have to compare against the realistic alternatives. For a Year 7–9 child at a UK independent school, the typical extra-tutoring options are:

Option Typical cost (per academic year) What you get
Nothing — school only £0 The school's teaching alone
Weekly human tutor, one subject £1,500–£3,000 30–40 hours of one-to-one human teaching
Twice-weekly human tutor across subjects £4,000–£7,000 Hours across multiple subjects
Specialist scholarship coach £3,000–£6,000 30–50 hours of paper-specific coaching
Online subscription tutoring (e.g. Sparx, MyMaths, Tassomai) £30–£200 Self-service adaptive practice
Self-service AI practice (aitutors.me — free demo, then £14 one-time top-up) £14–£56 Conversational, Socratic, available daily
Coach-led AI tuition (aitutors.me PRO, one subject) £1,080 (£360/term × 3) Named human coach + full professor faculty

The AI tutor is in a different price bracket from human tutoring by an order of magnitude. But it is also offering something genuinely different from the online subscription services.

What a year of it actually buys you, concretely

For aitutors.me specifically (your mileage will vary with other providers):

  • Socratic maths tutoring in KS3 (Years 7–9), available every day.
  • Energy gating — daily check-in that adapts session intensity to the child's state.
  • Parent dashboard — visibility into what was studied, for how long, at what energy.
  • Per-session retention controls — parent chooses what's kept, what's deleted.
  • Mentor + Professor Pi — two tutors at launch; more by faculty vote.

Functionally similar to ~20–25 hours of one-to-one human tutoring per year at a quality bar that's not the equal of a great human tutor but is the equal of an average one, available whenever the child needs it.

A rough but honest comparison: a month of AI practice covers the same cognitive ground as a £40/hour human tutor seen once a week for half a term. That's the calibration.

When the AI tutor genuinely substitutes for a human tutor

Three specific cases.

Case 1: Daily fluency, not specialist coaching

If your child needs more reps on KS3 maths topics they already understand the principles of, an AI tutor is not just cheaper — it's structurally better-suited. A human tutor at £60/hour for fluency reps is over-engineered.

Case 2: The 9pm prep crisis

The case the school doesn't see and the human tutor can't be present for. AI tutoring substitutes for nothing because the alternative is the child being stuck without help. This is the highest-ROI use case bar none.

Case 3: Topics the child won't ask about in class

The quiet child who won't admit confusion in a class of 14. AI tutoring covers a real and otherwise-invisible gap. The human tutor at £60/hour, seen weekly, also addresses this — but at vastly higher cost.

When the AI tutor doesn't substitute for a human tutor

Equally three specific cases.

Case 1: Scholarship-paper coaching

A specialist who knows the Westminster academic scholarship paper, the King's Scholarship paper, or the St Paul's entrance paper — and who has seen dozens of recent candidates — is delivering something an AI tutor cannot. Worth £80–£150/hour for the right child at the right stage.

Case 2: A child who needs human presence to engage

Some children — including some very academically capable ones — work better with another human in the room. Not all children are equally suited to text-based tutoring. We've seen children who thrive on it and others who don't. Try it. If after three weeks the child is consistently disengaged, the answer is human.

Case 3: Subjects we don't yet cover

aitutors.me at MVP covers KS3 maths. If your child needs help with English Literature, the Sciences, French, Latin, Spanish, or History, we are not yet the answer. (Some of these are on the roadmap; check the Faculty section on the homepage.)

A worked example — the typical Year 8 family

Let's take a Year 8 family at a competitive London day school. Annual education budget for school + extras: ~£25,000 + ~£3,000 = £28,000.

Within the £3,000 of extras, a reasonable composition:

  • £1,500/year — weekly human maths/English tutor (for scholarship prep, Saturday morning, one-to-one)
  • £600/year — half-termly intensive (school holidays, focused subject coaching)
  • £200/year — exam practice papers and supplemental materials
  • ~£150/year — AI tutoring (aitutors.me self-service practice, evenings)
  • £200/year — music exam fees, piano lesson top-ups, etc.
  • £350/year — sport, art, drama clubs not included in fees

The AI tutoring is the smallest line item. It is also, by design, the line item built to drive the largest day-to-day habit change in the child. The human tutor on Saturday is more academically valuable per hour; the AI tutor on Tuesday-Wednesday-Thursday is designed to carry the cumulative effect.

Where the AI tutor adds zero value

Some children and some weeks don't benefit. Be honest if you're in one of these cases:

  • The child is overscheduled. If prep is already finishing at 10pm because of activities, adding another tool — even a good one — is unlikely to help. The right answer is fewer activities.
  • Screen time is already high. A child with 4+ hours of recreational screen time daily does not need more screen exposure, even productive screen exposure. Reduce the entertainment use first.
  • The child has specific learning needs. Dyslexia, ADHD, dyscalculia, autism — these benefit from human specialist support that AI tutors do not currently match.
  • The child is academically very secure. A child consistently in the top quartile in maths, doing fine in school, with confident parents — may simply not need extra tutoring in that subject.

In each of these cases, the honest answer is "we're not what you need." We say so on the homepage. We mean it.

The asymmetry that drives the answer

If AI tutoring works for your child, it's one of the highest-ROI lines in your education budget. If it doesn't, the free demo cost you nothing and a £14 top-up is a single month — no contract, no retention call.

The asymmetry favours trying. The downside is small and reversible; the upside, when it works, is genuinely meaningful.

A final note on trying it cheaply

The honest way in costs nothing: run the free live demo first — no card, no signup. If it clicks, a £14 one-time Extra Time top-up buys a month of daily practice, the price of a single human tutoring hour. Step up to coach-led PRO tuition (£360/term per subject) only once you've watched your child actually use it. Current tiers are always at /pricing.

That's the honest ROI. The next move is yours.

FAQ

How much does a private tutor cost in the UK in 2026?

Typical rates: £40–£60/hour for general subject tutoring, £60–£80/hour for 11+/13+ specialists, £80–£150/hour for Oxbridge interview coaching. London rates run at the upper end of these bands.

How do I try AI tutoring without committing much?

Start with the free live demo — no card, no signup. If it clicks, a £14 one-time Extra Time top-up buys a month of practice. Coach-led PRO tuition (a named human coach plus the full professor faculty) is £360/term per subject. Current tiers are at /pricing.

When would I recommend NOT using AI tutoring?

For very young children (Years 3–5), for children with specific learning needs that benefit from human face-to-face interaction, and when family screen time is already high and adding more is counterproductive.

Duke Harewood runs aitutors.me. He is genuinely worse at maths than his daughter and grateful that Professor Pi exists. Updated 21 May 2026.