Codex
Claude
GeminiWhat you do
- Define the button choices and what each choice is allowed to do.
- Review any choice that can create work or send a material response.
- Inspect the stored outcome when follow-up is complete.
Teams using interactive Telegram messages / communications
Telegram Callback Follow-Up
A context-aware follow-up that acknowledges the interaction and performs only the approved next action.
From request to resultA complete plan you can follow.
Visible planThe problem
A button click can lose its original context or trigger an answer that overstates what happened.
AI can interpret the stored interaction alongside the original request while provider tools remain limited to explicit acknowledgements and sends.
What the agent does
- Load the original message, interaction record, and allowed callback policy.
- Resolve the selected choice without treating callback data as an authoritative instruction by itself.
- Acknowledge promptly, complete only the authorized work, and record the result.
When this workflow helps
A reusable method, adapted to the request.
- A Telegram callback_query created an agent_request.
- The agent needs to verify stored button state before acting on callback_data.
Before it starts
Connection required
- A connected Telegram profile
- Stored interaction context for each button
- A clear policy for what each choice authorizes
What proves it worked
Evidence, not a success claim.
- Callback acknowledgement receipt
- Linked interaction and source-message refs
- Recorded downstream outcome or blocker
Workflow path
The reusable stages of the work.
A real run expands these stages around the request, context, selected tools, approvals, and dependencies.
- 01Passes to next stage
Read Callback Context
Load the inbound callback, outbound button, and originating request metadata.
- 02Passes to next stage
Acknowledge Callback
Optionally clear Telegram client loading state with static acknowledgement text.
- 03Verified outcome
Send Follow-Up When Needed
Send an origin-bound reply only if the agent decision requires it.
Safe stopping and recovery
Useful even when the whole path cannot run.
- Stop after a safe acknowledgement when the selected action lacks context or approval.
- If interaction context is missing or expired, acknowledge safely and request a fresh choice.
- If downstream work is blocked, record the blocker without claiming completion.
Specialists inside this workflow
Clear roles for each part of the job.
Connected work