By Sagar Shankaran, Founder of CallSphere
Master timezone detection, locale-aware date formatting, and cross-timezone scheduling in AI agents to deliver accurate, localized time information to users worldwide.
Key takeaways
When an AI agent tells a user "your appointment is at 3 PM," the natural follow-up question is: 3 PM where? Global agents must resolve ambiguous time references, convert between zones, and present dates in the format the user expects. Getting this wrong causes missed meetings, incorrect data analysis, and eroded trust.
The core complexity comes from three sources: timezone offset is not fixed (daylight saving time changes it), date format conventions vary by locale (MM/DD vs DD/MM), and natural language time references ("next Tuesday," "tomorrow morning") depend on the user's local time, not the server's.
Store all timestamps in UTC internally and convert only at the presentation layer. Attach the user's timezone to their session context.
Hear it before you finish reading
Talk to a live CallSphere AI voice agent in your browser — 60 seconds, no signup.
flowchart LR
INPUT(["User intent"])
PARSE["Parse plus<br/>classify"]
PLAN["Plan and tool<br/>selection"]
AGENT["Agent loop<br/>LLM plus tools"]
GUARD{"Guardrails<br/>and policy"}
EXEC["Execute and<br/>verify result"]
OBS[("Trace and metrics")]
OUT(["Outcome plus<br/>next action"])
INPUT --> PARSE --> PLAN --> AGENT --> GUARD
GUARD -->|Pass| EXEC --> OUT
GUARD -->|Fail| AGENT
AGENT --> OBS
style AGENT fill:#4f46e5,stroke:#4338ca,color:#fff
style GUARD fill:#f59e0b,stroke:#d97706,color:#1f2937
style OBS fill:#ede9fe,stroke:#7c3aed,color:#1e1b4b
style OUT fill:#059669,stroke:#047857,color:#fff
from dataclasses import dataclass
from datetime import datetime, timezone
from zoneinfo import ZoneInfo
from typing import Optional
@dataclass
class UserTimezoneContext:
timezone_name: str # e.g., "America/New_York"
locale: str = "en-US"
@property
def tz(self) -> ZoneInfo:
return ZoneInfo(self.timezone_name)
def now(self) -> datetime:
"""Current time in the user's timezone."""
return datetime.now(timezone.utc).astimezone(self.tz)
def to_user_time(self, utc_dt: datetime) -> datetime:
"""Convert a UTC datetime to the user's local time."""
if utc_dt.tzinfo is None:
utc_dt = utc_dt.replace(tzinfo=timezone.utc)
return utc_dt.astimezone(self.tz)
def to_utc(self, local_dt: datetime) -> datetime:
"""Convert a user's local datetime to UTC."""
if local_dt.tzinfo is None:
local_dt = local_dt.replace(tzinfo=self.tz)
return local_dt.astimezone(timezone.utc)
Timezone detection typically relies on client-side JavaScript sending the Intl timezone, or IP-based geolocation as a fallback.
import httpx
from typing import Optional
class TimezoneDetector:
async def from_ip(self, ip_address: str) -> Optional[str]:
"""Detect timezone from IP using a geolocation API."""
try:
async with httpx.AsyncClient() as client:
resp = await client.get(
f"http://ip-api.com/json/{ip_address}",
params={"fields": "timezone,status"},
)
data = resp.json()
if data.get("status") == "success":
return data.get("timezone")
except Exception:
pass
return None
def from_utc_offset(self, offset_minutes: int) -> str:
"""Map a UTC offset to a common timezone (imprecise but useful as fallback)."""
offset_map = {
-480: "America/Los_Angeles",
-420: "America/Denver",
-360: "America/Chicago",
-300: "America/New_York",
0: "Europe/London",
60: "Europe/Paris",
330: "Asia/Kolkata",
540: "Asia/Tokyo",
600: "Australia/Sydney",
}
return offset_map.get(offset_minutes, "UTC")
Different locales expect different date formats. Build a formatter that respects the user's conventions.
from babel.dates import format_datetime, format_date, format_time
from datetime import datetime
class LocaleDateFormatter:
def __init__(self, locale: str = "en_US", tz_name: str = "UTC"):
self.locale = locale
self.tz_name = tz_name
def format_full(self, dt: datetime) -> str:
"""Format datetime with full locale conventions."""
return format_datetime(dt, format="long", locale=self.locale, tzinfo=ZoneInfo(self.tz_name))
def format_short_date(self, dt: datetime) -> str:
return format_date(dt, format="short", locale=self.locale)
def format_time_only(self, dt: datetime) -> str:
return format_time(dt, format="short", locale=self.locale, tzinfo=ZoneInfo(self.tz_name))
def format_relative(self, dt: datetime, now: datetime) -> str:
"""Human-readable relative time like 'in 2 hours' or '3 days ago'."""
diff = dt - now
seconds = diff.total_seconds()
if abs(seconds) < 60:
return "just now"
minutes = int(seconds / 60)
if abs(minutes) < 60:
return f"in {minutes} minutes" if minutes > 0 else f"{abs(minutes)} minutes ago"
hours = int(minutes / 60)
if abs(hours) < 24:
return f"in {hours} hours" if hours > 0 else f"{abs(hours)} hours ago"
days = int(hours / 24)
return f"in {days} days" if days > 0 else f"{abs(days)} days ago"
When an agent schedules a meeting between users in different timezones, present the time in each participant's local zone.
from dataclasses import dataclass
from typing import List
@dataclass
class Participant:
name: str
timezone: str
def format_meeting_for_participants(
utc_time: datetime, participants: List[Participant], formatter_locale: str = "en_US"
) -> dict:
"""Show meeting time in each participant's local timezone."""
result = {"utc": utc_time.isoformat(), "local_times": []}
for p in participants:
tz = ZoneInfo(p.timezone)
local = utc_time.astimezone(tz)
fmt = LocaleDateFormatter(locale=formatter_locale, tz_name=p.timezone)
result["local_times"].append({
"participant": p.name,
"timezone": p.timezone,
"local_time": fmt.format_full(local),
})
return result
# Usage
meeting_utc = datetime(2026, 3, 20, 14, 0, tzinfo=timezone.utc)
participants = [
Participant("Alice", "America/New_York"),
Participant("Kenji", "Asia/Tokyo"),
Participant("Priya", "Asia/Kolkata"),
]
schedule = format_meeting_for_participants(meeting_utc, participants)
Store it. Timezone detection from IP is imprecise and adds latency. Let users set their timezone explicitly during onboarding, detect it via JavaScript as a default, and allow them to change it in settings. Store the IANA timezone name (like "America/Chicago"), not a raw UTC offset.
Still reading? Stop comparing — try CallSphere live.
CallSphere ships complete AI voice agents per industry — 14 tools for healthcare, 10 agents for real estate, 4 specialists for salons. See how it actually handles a call before you book a demo.
Parse relative date references using the user's local time, not the server's UTC clock. Libraries like dateparser or python-dateutil can parse natural language dates. Always confirm with the user by echoing back the resolved date in their local format before scheduling anything.
Always use IANA timezone names (ZoneInfo) rather than fixed offsets. The zoneinfo module in Python 3.9+ handles DST transitions automatically. Never store or compute with raw offset values like UTC-5, because that offset changes when DST begins or ends.
#TimezoneHandling #DateFormatting #Globalization #AIAgents #Scheduling #AgenticAI #LearnAI #AIEngineering

Written by
Sagar Shankaran· Founder, CallSphere
LinkedInSagar Shankaran is the founder of CallSphere, where he builds production AI voice and chat agents deployed across healthcare, hospitality, real estate, and home services. He writes about agentic AI, LLM engineering, and shipping voice agents that handle real calls in production.
See how AI voice agents work for your industry. Live demo available -- no signup required.
A founder's guide to the personal AI assistant market: best AI assistant apps, business-grade options, and how CallSphere's voice agent fits in.
A founder's guide to free AI agents, low-code AI agent builders, and how to know when you should pay for a real platform like CallSphere.
Graphiti is the open-source temporal knowledge graph for AI agents in 2026. Learn how bi-temporal memory beats vector RAG for voice agents and long-running LLMs.
Chatbot app vs ChatGPT in 2026: a founder's clear take on the difference, when to use which, and how a real AI chatbot app development works.
How we built a fault-tolerant HVAC emergency triage and tech-dispatch platform on Kubernetes — three-tier CQRS, 11 micro-agents on the OpenAI Agents SDK + LangGraph, NATS JetStream, DTMF/SMS/WebSocket acceptance, circuit breakers, and an evaluation pipeline that catches regressions before they wake a tech at 3 AM.
Head-to-head: OpenAI Frontier and Anthropic's managed agent stack — strengths, fit, and what each means for enterprise AI voice and chat deployment.
© 2026 CallSphere Inc. All rights reserved.
Made within San Francisco