Tally
Tally is a schema-driven study assistant designed for CIMA Management Accounting students. It behaves as a governed agent — not a general-purpose chatbot — with explicit workflows and output contracts that keep study support structured, consistent, and exam-safe.
Who Tally works for
- CIMA students preparing for Management Accounting assessments
- Certificate, Operational, Management, or Strategic levels (depending on your context)
Tally helps learners plan, learn, practise, and review progress with structured outputs designed for exam readiness.
Tally is a governed educational assistant with explicit rules, contracts, and workflows. It does not improvise authority or advisory scope: it stays within exam-focused learning support.
When uncertain or missing inputs, Tally asks for what it needs rather than guessing.
What Tally is designed to do
- Onboard you with exam-relevant constraints (level, papers, dates, hours available)
- Create structured, time-aware study plans with milestones
- Teach syllabus-aligned topics with exam-focused explanations
- Run synthetic drills and mock exams (non-proprietary)
- Analyse performance, identify weaknesses, and adapt the plan
- Generate short-horizon booster plans leading up to exam day
- Export and re-import study context for continuity across sessions
How Tally works
Each message is processed through a controlled lifecycle: parse → route → govern → execute workflow → output.
1. Input parsing
Detects commands or infers intent, then maps your request to a supported workflow.
2. Safety and exam integrity
Applies privacy, exam-safety, and scope rules before producing any output.
3. Workflow execution
Runs a defined workflow such as onboarding, planning, topic help, drills, mocks, or progress review.
4. Structured output contracts
Produces predictable artefacts (plans, quizzes, mock reports, progress summaries) in consistent formats.
5. Quality gating
Applies clarity and scope checks; flags uncertainty and refuses unsafe requests where needed.
6. Portable continuity
Long-term continuity is user-controlled via export/import snapshots rather than hidden memory.
What Tally can work with
- Natural-language study requests and explicit slash commands
- Your CIMA level, target papers, exam dates, and weekly study time
- Study activity logs and self-ratings
- Drill and mock performance results
- Previously exported Tally context snapshots
Privacy-safe by default
Tally does not require real names or personal identifiers. It operates within your ChatGPT session and supports privacy-safe context exports for continuity.
Outputs you can expect
- Multi-week study plans with milestones and revision cycles
- Topic teaching sessions with worked, synthetic examples
- Targeted drills and mock-style practice with feedback
- Performance analysis and weakness maps
- Pre-exam booster plans and last-minute revision schedules
- Optional context exports to continue later without losing progress
Key commands
- /help — show available commands and workflows
- /tally_status — show current study status and next suggestions
- /onboard (or /start) — onboarding and profile setup
- /new_plan (or /study_plan) — generate a structured study plan
- /topic_help — teaching and practice for a specific topic
- /drill_weak_topics — drills focused on weaker areas
- /quick_drill — short, time-boxed drill
- /mock_exam — sit a mock exam
- /mini_mock — shorter mock format
- /log_study — log completed study activity
- /review_progress — analyse performance and adapt plan
- /exam_booster (or /last_minute_plan) — pre-exam booster schedule
For longer work across sessions, use /export_context and /import_context. See How to use AgentVon.
What Tally is not
- A source of proprietary CIMA exam questions or confidential materials
- An investment, tax, legal, or professional finance adviser
- An employer tool, institutional platform, or awarding-body representative
- A general-purpose assistant for unrelated topics
Try Tally, then imagine governed workflows elsewhere
Tally demonstrates structured, exam-safe support through defined workflows and predictable outputs. The same discipline can be applied to professional workflows where scope and quality matter.