Every rating requires cited evidence from the lesson observed, hedged feedback is flagged and replaced with a specific action, and the teacher holds a documented right of reply.
Why most observation changes nothing
The familiar pattern: a formal observation once or twice a year, a form completed afterwards, a paragraph that says the lesson was good with a target to develop questioning. The teacher plans something they would not normally teach. Everyone knows. Nothing changes.
Three specific faults, all fixable:
- Too infrequent. Two observations a year produce anecdotes. Ten short ones produce a trend and, more importantly, a working relationship in which being observed is ordinary.
- Too vague. “Develop questioning” gives a teacher nothing to do on Monday.
- Too entangled with judgement. Where observation feeds directly into performance management, teachers perform. Honest development becomes impossible.
How Edves does it
Evidence is required
An observer cannot save a rating without attaching evidence from the lesson — what was said, what pupils did, what was on the board or in books. A rating with no evidence does not save. This single constraint is the difference between an observation record and a form.
Hedging is flagged
Draft feedback is checked for the language that makes observation useless — “try to”, “generally good but”, “perhaps consider” — and the observer is prompted to state the specific action instead.
Short and frequent
Built for learning walks and brief drop-ins conducted often rather than formal annual observations. Also more realistic for a middle leader carrying a teaching timetable.
Right of reply
Every record carries a teacher response field that becomes part of the record. Not a courtesy: it is what makes teachers willing to be observed honestly, and what makes a record defensible if disputed — including in a union conversation.
Coaching linked to the finding
Each rubric area links to a next step and short development the teacher can act on before the next lesson, rather than a twilight session in six weeks.
CPD records
A continuous record per teacher: sessions completed, observation history, coaching cycles, internally delivered development, and external courses. When appraisal comes round, or a teacher moves school, or a school needs to evidence professional development under leadership and governance, the record already exists.
For early career teachers, induction and mentoring records sit alongside the same observation history, so support is based on what has actually been seen rather than a general impression.
What leadership gets
- Practice trends by teacher, department, key stage and school.
- Observer consistency. Where two observers score comparable lessons differently, the variance is surfaced. In most schools this is the first thing worth fixing and almost never examined.
- Cycle coverage — who has been observed and who has not.
- Link to planning and outcomes, connecting observation to the lessons and curriculum objectives from Pedagogy with AI and to assessment.
- Workload evidence — what CPD was delivered, and what practices were stopped.
Where AI is and is not used
Stating this plainly, because staff and unions will ask:
- AI drafts feedback language, surfaces relevant rubric areas from the observer’s evidence notes, flags hedged phrasing and suggests development resources.
- AI does not assign ratings, does not make employment or capability recommendations, and does not observe.
- A human observer is in the room and a human observer scores.
Schools introducing AI-supported observation should say this explicitly and in writing before the first observation. Ambiguity here destroys trust in the whole exercise faster than anything else.
Sequencing
| Stage | Focus |
|---|---|
| 1 | Load the school’s existing observation form. Do not change the instrument first |
| 2 | Calibrate observers — several score the same lesson and compare |
| 3 | Short learning walks only, framed explicitly as developmental |
| 4 | Introduce trend reporting once there is enough data to mean something |
| 5 | Connect to appraisal only once staff trust the developmental use |
Schools that begin at stage five get compliance and no change in practice.