How Indian Phone Numbers Should Be Cleaned Before a Campaign

Indian phone numbers should be cleaned before a campaign by selecting India as the country context, removing harmless presentation characters, interpreting recognised prefixes consistently, rejecting incomplete or structurally invalid values, and reviewing duplicates after normalisation. Cleaning makes records suitable for deterministic dialling; it does not prove that a number is active, belongs to the named person, has consented, or may lawfully be called. MYLO can guide the preparation of a supported lead file and show which records are eligible or rejected under its configured rules. The customer remains responsible for the source, consent and preference evidence, suppression lists, contact purpose, permitted times, provider requirements, and current Indian law and regulation. A normalised number is a technical input, not legal approval.

Select the Country Context First

A local-looking value is ambiguous without a country. Selecting India tells the preparation process how to interpret familiar domestic and international prefixes and what structural expectations to apply. Do not infer the country merely from the business address when a file may contain international customers.

If the audience spans several countries, separate the records or provide a reliable country field under a supported workflow. One blanket transformation can corrupt valid foreign numbers. The user should see the selected context in the preview and approval summary.

Remove Formatting, Not Meaning

Business files often contain spaces, hyphens, brackets, or text-formatting artefacts. A controlled cleaning step can remove characters used only for display while preserving the actual digits and recognised leading plus. It can also trim surrounding whitespace introduced by copying data between systems.

Avoid spreadsheet conversion that turns a long phone value into scientific notation, removes a leading zero, or rounds digits. Phone numbers are identifiers, not quantities. Store and exchange them as text. If the original value has already been damaged, reject or correct it from an authoritative source rather than inventing the missing digits.

Handle Prefixes Consistently

The same Indian destination may appear in domestic form, with a leading zero, with country code 91, or with +91. Normalisation should interpret only the prefix forms supported by the current product and produce one consistent representation for comparison and dialling. The provider may require a particular outbound format, which belongs in the trunk and dialling design rather than in ad hoc spreadsheet edits.

Do not strip 91 from every value that begins with those digits without considering length and country context. Likewise, do not prepend +91 to every numeric cell blindly. A safe process validates relationships between prefix and remaining digits and reports ambiguous records for review.

Reject Values That Cannot Be Safely Normalised

Blank cells, alphabetic placeholders, extensions embedded in the phone field, obviously incomplete values, and unsupported lengths should not enter the eligible campaign set. Report a reason for each rejected category. A user can then correct source data or knowingly exclude it.

Structural rules should be described as preparation checks, not as a complete statement of Indian numbering regulation. Numbering plans and provider practices can change, and different service types may follow different patterns. Current authoritative guidance and the customer’s telecom provider remain important for specialised cases.

Deduplicate After Normalisation

Two cells can look different but represent the same destination, such as a domestic form and a +91 form. Deduplicating raw text will miss that relationship. Compare the canonical prepared values, then show the business which records collide.

A shared family or office number may legitimately appear against several people. The system should not guess which record to keep or call the shared number repeatedly. The campaign owner must decide whether to merge the business context, select one record, or exclude the number. Suppression should also apply to the canonical destination across formatting variants.

Preserve Useful Audit Evidence

The preview should report total source rows, values accepted unchanged, values normalised, invalid values, blanks, and duplicates. For authorised reviewers, it can show a limited sample of original and prepared forms so the transformation is understandable. Avoid exposing the complete audience in general logs or screenshots.

Keep enough lineage to identify the approved source and preparation result. If a rule or source file changes, create a new review result rather than silently rewriting the campaign’s input. Protect both original and prepared data under the customer’s access and retention policy.

Practical Cleaning Example

A service company exports 2,100 reminder records. The phone column contains spaces, hyphens, domestic prefixes, +91 forms, blank cells, and a few spreadsheet-damaged values. The campaign owner selects India, verifies the phone column, and runs preparation. MYLO groups records into eligible, invalid, and duplicate categories without starting any calls.

The owner corrects damaged values from the company’s authoritative customer system, removes contacts covered by current suppression, and resolves shared-number duplicates. A new preview shows the exact eligible count. The team approves that version, then separately configures campaign flow, authorised caller IDs, schedule, pacing, retry, recording choice, capacity, and functional testing.

Consent, Preferences, and Regulatory Responsibility

Number formatting says nothing about whether contact is permitted. The customer must apply current Indian legal, regulatory, contractual, and industry requirements relevant to its organisation and campaign type. That may include consent or preference evidence, do-not-contact or suppression obligations, permitted purposes and time windows, sender or caller identity rules, and provider terms.

MYLO should not label a record “legally valid” simply because its digits match a structural rule. Legal review may be necessary, particularly for marketing, financial, healthcare, political, or other sensitive campaigns. A business should maintain a process for objections and ensure that a new file cannot reintroduce a suppressed destination.

Separate Data Preparation From Live Dialling

Cleaning and preview are guided preparation stages. After approval, deterministic campaign services manage eligible lead claiming, pacing, capacity, attempt state, and recovery. The LLM may help explain errors and collect requirements, but it should not improvise per-lead transformations or live retry decisions.

The provider ultimately accepts or rejects the dialled address under the configured trunk. A technically prepared value can still fail because it is disconnected, unreachable, barred, or not accepted by the provider. Campaign reporting should preserve that observed outcome without changing the underlying contact permission.

Indian Number Preparation Checklist

• Confirm the authorised source, purpose, audience, and accountable owner.
• Select India explicitly and separate genuine international records.
• Keep phone identifiers as text to avoid spreadsheet damage.
• Remove only harmless formatting characters.
• Interpret supported domestic and country-code prefixes consistently.
• Reject blanks, ambiguous values, and unsupported lengths with reasons.
• Deduplicate and suppress using the canonical prepared destination.
• Review limited original-to-prepared samples and category counts.
• Do not treat structural validity as consent, ownership, or legal approval.
• Approve the exact preparation result before campaign configuration and testing.

Good normalisation gives deterministic dialling a consistent input while preserving human accountability. When a value is ambiguous, stop and review it; a guessed digit can contact the wrong person.

Test the Prepared Values Against the Real Dialling Path

A preparation rule can produce a consistent canonical number while the provider expects a different outbound presentation. Use a small, authorised test cohort to verify the complete relationship between stored destination, dialled address, provider acceptance, and observed result. Keep provider-specific rendering inside controlled telephony configuration rather than rewriting the business source differently for every trunk.

Test representative input forms: domestic, country-coded, values with presentation characters, and records near structural boundaries. Confirm that each valid source maps to the expected canonical destination and that invalid inputs remain rejected. Do not test by calling arbitrary numbers; use controlled destinations with permission.

Wrong-number and reassignment handling

A structurally valid phone may now belong to someone else. Give agents and operators a clear way to record wrong-number or do-not-contact outcomes. Apply the resulting suppression to the canonical destination across all pending campaigns, not only the row where it was discovered. Do not “correct” ownership based on an agent’s guess.

Change control for preparation rules

If the supported normalisation rules change, existing approved campaigns should not be silently reinterpreted. Create a new preparation result, compare category counts, review material differences, and obtain approval again. Versioned evidence prevents later disputes about which digits were actually approved for a campaign.

Operator Review Questions

Before approval, ask whether any values unexpectedly changed country context, whether large groups failed the same rule, whether duplicates cluster around shared office or household numbers, and whether the eligible count matches business expectations. A sudden drop may reveal the wrong column or spreadsheet damage rather than a genuinely poor list.

Confirm that all current suppression sources were applied after normalisation and that the review sample does not expose unnecessary personal data. The reviewer should be able to explain every category count and identify the authoritative source for corrections. If not, return the file for preparation rather than accepting uncertainty into live execution.