Sign In
Try Now

AI Insights

AI Insights turns every kudos your team sends in Slack into organized, interpretable data about your culture. Instead of counting recognition, AsanteBot’s AI reads it — mapping each message to a company value, tagging it by recognition category, and surfacing patterns across teams, trends, and time periods.

This guide walks through everything you need to know: what AI Insights does, how to set it up, and how to use each screen effectively.

Table of contents

What is AI Insights?

AI Insights is a built-in analytics layer inside AsanteBot that automatically classifies every recognition message in your Slack workspace against two things:

  • Company values — the cultural themes your organization cares about (e.g. “Celebration of Success,” “Collaboration and Connectivity,” “Continuous Improvement”).
  • Recognition categories — the types of contributions being recognized (e.g. “Everyday Appreciation,” “Celebrations & Milestones,” “Rewards Earned”).

Once enabled, AI Insights gives People teams, managers, and leadership three key views:

  • An Overview dashboard showing value distribution, an organizational network map, and recognition category breakdown.
  • A Feed Analysis table showing every individual kudos with its AI-assigned value and category.
  • A Classify screen where you define and edit the labels the AI uses.

Before you start

To get the most out of AI Insights, make sure:

  • AsanteBot is installed in your Slack workspace and your team has started sending recognition.
  • You have admin access to the AsanteBot dashboard.
  • You have a public company URL ready (homepage, about page, or careers page) — the AI uses this to suggest values tailored to your organization.

Setup takes about 5 minutes end-to-end.

Step 1 — Enable AI Insights

Log in to the AsanteBot dashboard and head to AI Insights → Settings in the left sidebar.

On the Settings page, find the Enable AI Insights card at the top and toggle the status to Enabled. Changes apply workspace-wide immediately.

Once enabled, AI Insights will begin analyzing recognition messages for quality, themes, and trends. Insights become available across the Overview and Feed Analysis screens.

Tip: You can toggle AI Insights off at any time without losing your historical recognition data.

Step 2 — Run the Classify Wizard

Before the AI can tag recognition meaningfully, it needs to know what to tag against. The fastest way to set this up is the Classify Wizard.

Go to AI Insights → Classify and click the Classify wizard button in the top-right corner.

How the wizard works

The wizard is a 4-step flow. On Step 1, you’ll be asked for:

  • Source URLs — add up to 5 public links (homepage, about page, careers page, culture page, mission statement, etc.). Fewer links keep the step simple; 2–3 strong pages is usually enough.
  • Company name (optional) — useful if your brand shares a name with other companies, so the AI doesn’t mix up context.

Click Generate suggestions and the AI will read your pages and propose a starting set of company values and recognition categories tailored to your organization.

You’ll then move through the remaining wizard steps to review, refine, and confirm the suggestions before they go live.

Step 3 — Review and edit your values & categories

After the wizard completes, you’ll land on the main Classify page. This is where your recognition taxonomy lives, organized into two columns.

Company values

Company values are the cultural themes the AI maps each kudos against. The screenshot example shows three active values:

  • Celebration of Success — recognize wins of all sizes, not just major milestones.
  • Collaboration and Connectivity — reward behaviors that bring people together, strengthen relationships, and make teamwork smoother.
  • Continuous Improvement — recognize people who learn, iterate, and use feedback to make the product, process, or team better.

Each value has an Edit button (to rename or update the description) and a Delete button. You can also click + Value in the top-right to add a new one manually.

Recognition categories

Recognition categories describe how a contribution helped the team. Default examples include:

  • Celebrations & Milestones — birthdays, work anniversaries, launches, and personal milestones.
  • Everyday Appreciation — quick peer-to-peer thanks for helpful actions and day-to-day contributions.
  • Rewards Earned — recognitions tied to redeemable points and standout contributions.

As with values, you can edit, delete, or add new categories using the + Category button.

Important: Create values and categories before enabling deep AI analysis. A cleaner taxonomy up front produces dramatically better insights later.

Using the Overview dashboard

The AI Insights → Overview screen is your at-a-glance view of recognition across your workspace. Use the date picker at the top to filter any range (the default shows the current trailing period).

Company values

The top section shows how recognition maps to your company principles. You’ll see two views side by side:

  • Value Distribution bar chart — visualizes how often each value is being recognized.
  • Value table — lists each value with its raw count and percentage share of total recognition.

This is the fastest way to spot whether your team is living all of your values or concentrating around just one or two. If one value dominates at 70%+ while another sits at 0%, that’s a cultural signal worth acting on.

Organizational network map

The middle section visualizes how people connect through recognition in the selected period. Each avatar represents a person, and connections show who is recognizing whom.

You can filter the map by:

  • All values — narrow to a specific company value.
  • All categories — narrow to a specific recognition category.
  • All groups — filter by team or department (requires HRIS integration).

Use Clear filters to reset the view. The network map is especially useful for spotting silos (teams that only recognize within themselves) and bridge-builders (individuals who consistently recognize across departments).

Recognition categories

The bottom section mirrors the values view but for contribution types. You’ll see a pie chart showing category distribution and a table with counts and percentage shares.

This view answers a practical question: are we mostly doing quick peer thank-yous, or are we actually recognizing milestones and rewarding standout work? A healthy program usually shows a mix across categories.

Using Feed Analysis

The AI Insights → Feed Analysis screen is the raw, row-level view of every recognition message with its AI-assigned tags. Think of it as the audit trail behind the Overview dashboard.

Each row shows:

  • Person — who was recognized (and who sent it).
  • Company Value — the value the AI mapped the message to (or “–” if no clear match).
  • Category — the recognition category assigned.
  • Message — the original kudos text.
  • Date — when the recognition was sent.

You can filter the table by All values and All categories using the dropdowns in the top-right. Use the date picker to narrow to a specific period.

Exporting to Excel

Click the Export to Excel button in the top-right to download the current view as a spreadsheet. This is useful for:

  • Sharing recognition data with leadership outside the platform.
  • Building custom quarterly or annual reports.
  • Cross-referencing recognition data with performance or retention metrics.

Note: messages with “–” in the Company Value column are ones the AI couldn’t confidently map to any of your defined values. These often indicate very short messages (e.g. “thanks!”) or a gap in your values taxonomy.

Analyze & Reclassify — keeping your data fresh

Whenever you edit your values or categories in Classify, your historical recognition data still carries the old tags. To re-apply your updated taxonomy across historical data, use the Analyze & Reclassify feature.

Find it in AI Insights → Settings, in the middle card.

How it works

When you click Analyze & Reclassify now:

  1. A background job is queued.
  2. Existing recognitions are re-evaluated using your current company values and categories.
  3. Your Overview dashboard and Feed Analysis update automatically once the job completes.

The settings card shows two useful metrics:

  • Attempts left this month — you have a limited number of reclassification runs per month (e.g. 2 / 2).
  • Last run — the date your data was last reprocessed.

When to run it: run Analyze & Reclassify after bulk edits in Classify, after adding new values, or any time your dashboards feel out of sync with your current taxonomy. Running it once after every major change is ideal.

Best practices

Follow the recommended setup order

  1. Define labels in Classify first, so the AI knows how to tag incoming messages.
  2. Enable AI Insights in Settings.
  3. Let data accumulate for at least a few weeks before drawing conclusions.
  4. Run Analyze & Reclassify after any major taxonomy change.

Keep values simple and distinct

Three to five clear, non-overlapping values work far better than ten similar ones. If two values feel like they’d apply to the same kudos, merge them. The AI (and your team) will thank you.

Write value descriptions clearly

The description on each value card isn’t just for humans — the AI uses it as context when classifying messages. A vague description (“teamwork is important”) produces vague tagging. A specific one (“reward behaviors that bring people together across channels and locations, including helping others and cross-functional support”) produces precise tagging.

Review Feed Analysis monthly

Once a month, spot-check a page or two of Feed Analysis. Are messages being tagged the way you’d expect? If not, tune your value descriptions and run Analyze & Reclassify.

Watch for “–” rows

A high proportion of untagged messages usually means one of two things: your value taxonomy has a gap, or your team is sending a lot of very short messages. Either way, it’s actionable information.

Frequently asked questions

Do I need to enable AI Insights for my whole workspace, or can I test it first?

AI Insights is a workspace-wide setting — when you enable it in Settings, it applies to all recognition data immediately. However, you can toggle it off at any time without losing data.

What happens to my existing recognition data when I turn on AI Insights?

Historical recognition is automatically analyzed and tagged using your current values and categories. You’ll see dashboards populate shortly after setup completes.

Can I edit the values and categories the AI suggests?

Yes. Every value and category has an Edit button in the Classify screen. You can rename, rewrite descriptions, or delete any suggestion. You can also add your own manually using the + Value or + Category buttons.

How many times can I re-run the Classify Wizard?

You can run the Classify Wizard multiple times to pull fresh suggestions from your site. Each run generates new proposals you can merge with or replace your existing labels.

How often should I run Analyze & Reclassify?

Run it after any significant change to your values or categories. You have a limited number of runs per month (shown in Settings), so save them for meaningful taxonomy updates rather than minor edits.

Why do some messages show “–” under Company Value?

The AI assigns “–” when it can’t confidently map a message to any of your defined values. This usually happens with very short messages (“thanks!”), messages that don’t match your current taxonomy, or recognition sent before you defined your values. Running Analyze & Reclassify often resolves the last case.

Is our recognition data used to train AI models?

No. Your workspace data is processed to generate insights for your organization only and is not used to train shared AI models.

Do I need an HRIS integration to use AI Insights?

No. AI Insights works with or without HRIS. An HRIS connection unlocks team/department filtering in the Organizational Network Map, but all other features work independently.

Need help? Reach out through the Let’s Chat button in the bottom-right of any AsanteBot dashboard screen, or visit our Help Center.