Separate the subject from the delivery
A human AI mentor and a coding tutor may teach overlapping material. The useful distinction is the learning work they can support. A coding tutor may focus on reading programs, writing functions, debugging and building reliable habits. An AI mentor may also need to explain datasets, model behaviour, evaluation and limitations. Neither title settles the question. Ask each person to describe the first two sessions using the same learner brief. Compare the proposed activities and responsibilities rather than the job title or the sophistication of the software they mention.
Start with demonstrated tasks
Record things the learner has actually attempted: changing a variable, tracing a loop, reading a table, finding a bug or explaining an output. Keep the amount of support visible. Someone who completed a guided notebook has a different starting point from someone who can adapt it independently, even when the final screens look similar. Age matters for eligibility and supervision, while these tasks help define the technical starting point. Do not convert the list into a composite ability score. It is a conversation about where useful teaching can begin.
Make room for both foundations and AI ideas
An introductory project can explore classification, examples and errors without requiring a large codebase. A more demanding application may expose missing coding foundations quickly. Ask whether the teacher will pause to build those foundations or quietly complete the difficult parts. The 2026 CSTA description gives attention to reading, evaluating, modifying and debugging programs alongside writing them. Learning Prospect uses that as a useful question for a mentor, not as a claim that any proposed lesson is standards-aligned. A learner should have opportunities to inspect and change the system, not merely launch it.
Choose the smallest sufficient appointment
If the main barrier is basic programming, a coding teacher with clear instructional practice may be a practical starting option. If the learner can already build small programs and wants to investigate model behaviour, a mentor with relevant AI expertise may fit. A parent-built pathway can also work where an adult has time, teaching confidence and the ability to manage accounts and materials. Whichever route you consider, ask what it cannot supply. Revisit the choice after a bounded project rather than treating the initial title as a permanent category for the learner.
Sources and limitations
Editorial decision guidance, not an assessment of an individual learner or provider. No lessons were observed. Linked sources provide context, not endorsement of Learning Prospect.
Source review: 18 November 2026. Provider terms can change. Read our methodology.