Skip to main content

How Bonusly uses AI

Learn how Bonusly uses AI to make recognition and performance smarter, faster, and more actionable.

Updated today

We build with AI at the core, but never at the cost of connection. Bonusly uses intelligent insights to make recognition and performance smarter, faster, and more actionable.

But the goal isn’t automation for its own sake—it’s amplification. We use AI to help managers lead with clarity, teams connect with purpose, and employees feel more seen, not less.

Bonusly’s AI helps your entire organization:

  • Recognize better, faster, and more meaningfully

  • Coach and celebrate as the work unfolds

  • Spot patterns, signals, and stories in your culture data

  • Turn your recognition data into real insights

Here's an overview of how AI is incorporated into our platform and features:

Video processing

  1. Video thumbnail selection

    1. Uses OpenAI to automatically select the best thumbnail from video frames

    2. Evaluates criteria like open eyes, closed mouth, and image quality

  2. Transcription Text Creation

    1. Uses Assembly AI to convert microphone’s audio recording into text. Supports dozens of languages.

  3. Transcription Text Correction

    1. Uses OpenAI to fix speech-to-text transcription errors, specifically company names and people names.

    2. Evaluates and replaces mis-spelled or mis-transcribed words with boosted words.

  4. Video Bonus Reason Generation

    1. Uses OpenAI’s GPT-4o-mini model to generate recognition messages from video transcripts.

    2. Maintains tone and emotional content while creating engaging recognition text.

Recognition and Feedback

  1. Quick Thank You Generation

    1. Generates personalized thank-you responses in user's writing style. Thank you options are presented to the user as an option to respond to the recognition they’ve received.

    2. Creates 9 unique suggestions per request

  2. Feedback Text Creation

    1. AI helps turn raw feedback into better, more constructive feedback.

    2. Tailored to user context and situation

  3. AI-powered gratitude gauge (Late August, 2025)

    1. Deploys AI to improve the quality of recognition people post.

  4. Recognition & Feedback Triggers from 1:1 Transcriptions (Late August, 2025)

    1. We analyze transcribed 1:1 conversations and goal completion using AI to detect recognition or feedback moments worth acting on, and surface smart suggestions in context

Recaps and summarization

  1. Team Recognition Summarization.

    1. OpenAI is used to summarize impact and areas of improvement from feedback and recognition posts.

    2. The summary is displayed in the top-left corner of the manager’s team page.

  2. User Nominations Summarization

    1. OpenAI analyze the nominations that have been made on behalf of a specific user and summarizes the collective sentiment as to why the user should receive the award.

    2. Results are displayed on the Admin Nominations Details page.

  3. One-on-One Summary Generation

    1. AI generates meeting recaps, summaries, and follow-up ideas for 1:1 meetings based on notes and transcripts.

    2. Agenda topics are also summarized. Summaries for each are displayed in the right-hand drawer on the one-on-one page.

    3. Displayed as the Meeting Summary and Suggested Topics on the one-on-one page.

Insights

  1. AI-Powered Quality Insights

    1. Produces a Recognition Quality Score, powered by ai, that evaluates the richness of recognition messages. Rather than simply tracking how much recognition is happening like our current analytics, we aim to assess how meaningful it is.

Compliance and business verification (KYB)

In addition to recognition, feedback, and performance features, Bonusly also uses AI behind the scenes to help our team verify new customer accounts.

When a company signs up for Bonusly, we may run a “Know Your Business” (KYB) review to check for sanctions, fraud, or other compliance risks. To support this process, we use OpenAI models to gather, then analyze a limited set of business information (such as company name, domains, and key contacts) and generate a structured risk assessment. All of the data we gather for KYB verification is in the public domain already.

A few important notes about this KYB use of AI:

  • It is used only for business‑level screening, not to make employment or performance decisions about individual employees.

  • It relies on company‑level details and public or customer‑provided information; it does not use the content of your recognition posts or feedback for risk scoring.

  • AI output is used to help our team standardize and accelerate reviews; we maintain human oversight and the ability to override automated recommendations where appropriate.

You may occasionally see references to automated KYB decisions (for example, notes attributed to “BonuslyGPT”) in internal audit or approval records. These simply indicate that an AI‑assisted workflow contributed to the decision, with our compliance team retaining ultimate responsibility for outcomes.

Did this answer your question?