Agent Profiles

Agent Profiles give you a deep, focused view of how each AI agent is performing against its goals. This is where activity turns into accountability, showing not just what an agent is doing, but whether it’s delivering outcomes that matter.

Each profile is organized around three core areas:

• Topline performance metrics

• Objectives & Key Results (OKRs)

• Campaign execution and impact

Agent Profiles are centered on Objectives and Key Results (OKRs), so the entire page is locked to the agent’s OKR annual timeframe. This ensures performance is always evaluated consistently against annual goals, not short-term fluctuations.

Topline Metric Tiles

Each agent profile displays the same three topline metrics shown on the main dashboard, scoped specifically to that agent:

• Inbound Messages

• Daily Active Users

• Quality Response Index

Each metric includes:

• A trend comparison of the last 7 days vs. the previous 7 days

• Visual indicators to quickly spot improvements or declines

Objectives & Key Results (OKRs)

OKRs are the foundation of agent accountability in the Console. Each agent is measured against industry-benchmarked objectives that reflect real business outcomes, not arbitrary activity metrics.

These benchmarks are based on patterns and performance standards observed across similar organizations and verticals over the last 10 years.

Key Results

Overview

Every Key Result (KR) in your Agent Performance Console gets a score from 0 to 1 based on how close your agent is to hitting its target. This score determines the progress indicator you see next to each KR.

The Score Scale

Score	Status	What It Means
1.0 (Dark Green)	Smashing Success	You've exceeded the target: outperforming your peers.
0.7 (Light Green)	Success	You're hitting the benchmark. This is the goal.
0.5 (Yellow)	Partial Progress	You're making progress but haven't reached the benchmark yet.
0.3 (Orange)	Low Progress	Early-stage progress. Common early in the cycle for cumulative KRs.
0 (Red)	Not Achieved	No measurable progress toward this KR yet.

How We Calculate a KR Score

Each KR score is determined in two steps:

Step 1: We calculate your actual result as a percentage of the target. For example, if the target is 1,000 resolved conversations and you're at 700, that's 70% to target.

Step 2: That percentage is checked against score thresholds specific to that KR.

How We Set Targets and Thresholds

Targets aren't arbitrary. Here's how we build them:

Start with the benchmark. For each KR, we determine the expected median result for businesses in your vertical. That median value becomes the threshold for a score of 0.7 (Success). This means hitting 0.7 puts you at or above the middle of your peer group.

Set the stretch target. We work backwards from the median to set the full target (1.0) — a round number where the median is roughly 70% of the target. This is intentionally a stretch goal.

Fill in the lower thresholds. We apply the same logic downward to set the ranges for Partial Progress (0.5), Low Progress (0.3), and Not Achieved (0.0).

Understanding KR Progress

Not all KRs behave the same way. The Console accounts for this by visualizing progress appropriately:

Rate-based KRs Examples: Quality Response Index, Positive Feedback Rate

• These may fluctuate throughout the cycle

• Display a 7-day trend comparison, similar to topline metrics

Cumulative KRs Examples: Revenue-driving conversations, customer preference data collected

• These increase over time

• Display a progress bar toward the annual target

When projecting status for cumulative KRs, the Console accounts for seasonality and industry patterns.

See the Full Library of Key Results with definitions and examples here.

Key Takeaways

• 0.7 is the goal. It's not "barely passing" — it means you're performing at or above the median of your peer group.

• Targets are stretch goals by design. Hitting 1.0 means you're significantly outperforming similar businesses.

• Benchmarks are industry-specific. They're built from nearly a decade of real performance data and account for vertical, venue type, and seasonality.

• Scores may shift over time. As we refine our data and onboard more clients, thresholds may be adjusted to stay accurate and relevant.

Objectives

Current vs Forecast

You'll see two scores for each objective & key result: where you are today and where we project you'll land by the end of annual cycle.

• Current Status = how you're performing right now against your annual target.

• Forecasted Status = where we expect you'll finish the cycle based on your trajectory so far. This is based on the seasonality of your vertical.

Why both matter: Current status tells you the score today. The forecast tells you whether you need to adjust now or if you're on pace.

Two types of key results behave differently:

• Cumulative KRs (e.g., 10,000 conversations resolved without a human agent) — these build over time. Current status will naturally start low early in the cycle. But the forecast might already show Success if the pace is strong enough to hit the target by December. So don't worry about red in March — check the forecast.

• Rate KRs (e.g., 70% positive feedback rate) — these can fluctuate up or down at any point. For now, current and forecasted status are the same because we're not assuming the rate will shift.

Campaigns

Agent Campaigns are structured, AI-driven missions assigned to individual agents that align behavior with strategic business outcomes. They let you activate those focused missions based on what your business needs right now — and measure whether they’re delivering results. They are the 'why & what', guided by measurable Key Results.

Agent Profile displays Campaigns summarizing active and completed campaigns tied to that agent.

Campaign Details

Selecting Campaign Details opens a detailed view that includes:

• Campaign description and goals

• Journey workflows

• Linked Key Results

• Integrations and data sources

Campaigns show how objectives are being activated in practice and whether those efforts are delivering measurable results.

You can set up your own campaigns based on your business needs.

Agent Chat

In your Console you can chat with your AI agent about their performance as if you’re talking with your employee about their job. This is designed to specifically provide you feedback on how the agent is working relative to its objectives and key results.

Agent Chat supports a wide range of performance-focused questions, including:

• Summaries of agent performance and trends

• OKR and KR status explanations

• Revenue and conversion insights

• Topic-level performance and trends

• Recommendations for improvement

Sample questions include:

• How’s progress on my OKRs?

• What topics are trending?

• Where are we underperforming?

• How do you make me money?

• Why is this KR low?

• How can we improve this objective?

• What conversations generate the most revenue?

• How do you calculate this KR?

Agent Chat turns performance data into conversations—making it easier to understand what’s happening and what to do next.