Skip to main content

Is Polly Anonymous on Slack and Teams? Settings, Exports, and Bot Logs

Is Polly anonymous in Slack and Teams? Learn how creator settings, audit logs, and export reports impact voter identity compared to zero-knowledge polling.

InviziPoll team7 min readPublished

Short answer: Polly offers anonymous voting settings in Slack and Microsoft Teams, but it is not anonymous at the infrastructure or network layer. Workspace administrators, bot authorization logs, and poll creators have access to participation records, timestamps, and export data that can link who voted with specific answers.

What we say about Polly here was last verified on 2026-10-13 against Polly's official documentation, Slack App Directory specifications, and Microsoft Teams integration guides.

When you respond to a poll inside Slack or Microsoft Teams, your response passes through corporate chat APIs before reaching the polling vendor's database.

Understanding Polly's three anonymity tiers

Polly is one of the most widely adopted survey bots for workplace communication tools. When an author creates a poll in a Slack channel or Microsoft Teams chat, Polly provides three distinct anonymity levels:

1. Non-anonymous (default)

In standard polls, anonymity is turned off completely. Every voter name and avatar appears directly beneath their chosen option in the public channel message. Anyone in the channel can see who voted for which option, when they submitted it, and whether they changed their vote later.

2. Anonymous

When the creator toggles the Anonymous setting, Polly hides voter names from the shared channel card and from the creator web dashboard. The card shows aggregate vote counts and percentage bars without displaying usernames. However, on the backend, Polly logs submission metadata, including unique user tokens and exact timestamps.

3. Confidential (Semi-anonymous)

The Confidential tier creates widespread confusion in workplaces. In this setting, the public channel card looks identical to an anonymous poll, hiding voter names from colleagues. However, the poll creator and authorized workspace admins can view individual respondent names and their exact answers inside the private Polly dashboard. Employees who vote assuming full privacy are often shocked to learn their manager received an itemized breakdown of their answers.

Comparing Polly voting modes with end-to-end encryption

The table below contrasts identity capture, creator visibility, and export data across Polly settings against true zero-knowledge encryption:

Setting or ModeVoter Names Shown in ChannelCreator Sees Individual NamesExport Contains User IdentifiersAnonymity Guarantee
Polly Non-AnonymousYes (names and avatars visible to all)Yes (names, emails, timestamps)Full user profile, email, answer, timestampNone (fully public in channel)
Polly ConfidentialNo (aggregate counts shown in chat)Yes (creator receives named roster in dashboard)Full user profile, email, answer, timestampNone (administrative policy only)
Polly AnonymousNo (aggregate counts shown in chat)No (names masked in dashboard)Voter token, answer, precise timestampPolicy-based (anonymized rows in cloud database)
InviziPoll Zero-KnowledgeNo (no channels, no accounts)No (cryptographically impossible)Shuffled aggregate totals, undated textCryptographic (WebCrypto browser E2EE, ciphertext only)

What poll creators and workspace admins see in exports

When a poll concludes, Polly allows creators and enterprise workspace admins to download CSV or Excel reports. Understanding what appears in these exports clarifies the boundary between policy promises and technical privacy:

1. Non-anonymous and confidential exports

Exports from non-anonymous and confidential polls contain complete employee profiles. Each row includes the employee Slack Member ID or Teams User Principal Name, display name, registered work email address, submission timestamp down to the second, and their selected answer.

2. Anonymous exports and token tracking

In anonymous polls, Polly masks the display name and email address columns in the exported spreadsheet. However, the export still generates individual rows for each response. Each row contains an internal voter token, the chosen option, and the exact submission timestamp.

3. The live counter stalking risk in small team channels

In small department channels with 5 to 10 colleagues, live counter updates introduce an immediate privacy hazard. Polly updates the channel tally in real time as each person clicks a button. If a manager posts a poll asking "Are you satisfied with project priorities?" and watches Slack, they see the negative tally increase the second an employee finishes reading the message. Even if names are hidden, timing correlation exposes the voter.

Platform-level visibility: Slack and Microsoft Teams bot architecture

Beyond Polly's own application settings, polling inside workplace chat platforms introduces systemic privacy concerns inherent to enterprise messaging software:

1. Slack Event API and bot token scopes

To post messages and record button clicks, the Polly Slack app requires extensive workspace permissions (chat:write, commands, im:history, users:read). When you click a voting button in Slack, Slack's servers generate an interactive message payload sent via webhook to Polly's cloud infrastructure. This payload contains your Slack User ID (U12345678), the Channel ID, and the Action ID. Both Slack internal systems and Polly servers process this event in plaintext before determining whether to mask your name in the UI.

2. Enterprise audit logs and discovery APIs

Enterprise Slack (Enterprise Grid) and Microsoft Teams (Microsoft Purview) retain detailed administrative audit logs. IT administrators can inspect interaction payloads, API gateway logs, and outgoing webhooks. If compliance teams or IT executives conduct an internal investigation, server-level logs can reconstruct who interacted with specific bot messages.

3. Centralized plaintext cloud databases

Polly stores poll definitions, user directories, and survey responses in centralized cloud databases. While Polly enforces security standards such as SOC 2 and TLS encryption in transit, data is decrypted into plaintext in server memory to generate charts, calculate averages, and produce reports. Company administrators rely on Polly's internal access controls rather than mathematical encryption boundaries.

How InviziPoll eliminates workplace polling vulnerabilities

InviziPoll was built from the ground up to eliminate the architectural risks of chat-bot polling and centralized survey platforms:

  • Browser-Based WebCrypto Encryption: Every response is encrypted inside the respondent's web browser before transmission across the internet. The server receives only unreadable ciphertext blobs.
  • Strict Database Minimization: The InviziPoll responses table contains exactly three columns: id, poll_id, and encrypted_blob. The platform has no user profile tables, no tracking cookies, and no links between respondent credentials and stored votes.
  • The Phase 120 Reserve Pool: Unlike Polly's instant vote increment counters, InviziPoll protects voters against live tally stalking. Results remain completely sealed until the poll reaches its minimum group size plus a 3-response reserve pool (max(3, k) + 3, which equals 8 responses when the default threshold k = 5 is selected). While the poll is active, at least 3 responses stay unreleased in the reserve pool. New submissions release only in randomized subsets of 3, 4, or 5 chosen by a CSPRNG. When the poll closes, all remaining unreleased responses in the reserve pool release together in one final batch.
  • No Channel Profiling or Bot Permissions: InviziPoll does not require administrative Slack or Teams workspace bot permissions. Polls are distributed via secure web links, keeping workplace communication separate from confidential employee feedback.

When to use Polly and when to use encrypted polling

Polly is a convenient tool for quick team logistics, casual lunch votes, scheduling sync meetings, and agile retrospectives where colleagues feel safe speaking openly.

When leadership needs authentic input on sensitive topics, executive performance, restructuring anxiety, or psychological safety, relying on chat-bot polls exposes workers to timing correlation and administrative surveillance. True candor requires cryptographic guarantees.

Learn more on our Polly comparison page, check our analysis of Slack poll anonymity, review Microsoft Teams poll privacy, explore our pricing plans, or examine our anonymity architecture.

FAQ

Can a Slack administrator see who voted in an anonymous Polly?

Slack workspace administrators cannot see names directly on the Polly channel card. However, Polly logs interaction payloads containing Slack User IDs on its servers, and enterprise IT admins with access to audit logs or outgoing webhook integrations can inspect low-level network traffic to correlate votes.

What is the difference between Anonymous and Confidential in Polly?

In an Anonymous Polly, respondent names are hidden from both the channel and the creator dashboard. In a Confidential Polly, names are hidden from colleagues in the public channel, but the poll creator receives an itemized report displaying each employee name alongside their chosen answer.

Can a manager figure out my vote from live Polly updates?

Yes. Polly updates vote tallies in real time as participants click options. In small team channels, a manager watching the channel can easily correlate an incremental change in the vote count with the exact moment an employee views the message or taps their screen.

Does Polly store employee survey responses in plaintext?

Yes. Polly stores survey responses in cloud database records protected by standard TLS encryption in transit. However, responses are processed as plaintext on Polly application servers to calculate aggregates and export reports, meaning data is protected by administrative policy rather than client-side encryption.

How does InviziPoll prevent timing correlation compared to Polly?

InviziPoll seals results until the poll reaches its minimum group size plus a 3-response reserve pool. While open, at least 3 responses remain unreleased, and new arrivals release only in randomized batches of 3, 4, or 5, releasing all remaining responses when the poll closes.

Share this articleLinkedInEmail
No credit cardStart free trial