Turn support conversations into product, CX, and revenue signals leaders can act on.
Each diagnostic is built from 1,000 representative support conversations for a sector, processed through the ArNa Resolve pipeline. It tells you not just whether the bot works, but whether it is handling the right work, escalating the right risk, and surfacing the signals that matter.
How the ArNa Resolve pipeline turns support conversations into decision-ready intelligence. Click to enlarge.
Choose a sector diagnostic
Each diagnostic is built from 1,000 representative support conversations processed through the ArNa Resolve pipeline.
Could not load data. Please try again — if this is your first visit today, Azure may need a moment to warm up.
B2B SaaS Support Analysis
Based on 1,000 representative synthetic support conversations processed through the ArNa Resolve pipeline
1,000 queriesLive data
Executive Verdict
—
Main constraint
—
Highest-value move
—
Watch risk
—
How to read this report:Green shows work the bot resolves cleanly. Red shows queries that escalated to humans — sometimes correct design, sometimes capacity strain. Amber shows queries outside reliable bot handling. Blue highlights documentation gaps.
Self-Serve Resolution
—
Resolved without a human
Escalation Rate
—
Sent to a human
Edge Case Rate
—
Outside reliable bot handling
Out-of-Scope Rate
—
Queries outside the bot's scope (real docs gaps appear in Top Documentation Gaps below)
Workload mix
—
The donut shows where demand enters the operation, before any deflection, escalation, or edge-case handling.
Deflection Rate by Category
Operational Findings
What the operation should learn
Finding 01
Intent Distribution
—
—
Prioritise the top two intents first. They are carrying most of the queue.
Finding 02
Deflection Rate
—
—
Fix the weakest category first. That is where the next lift in deflection will come from.
Finding 03
Escalation Triggers
—
—
Review escalation clusters weekly. Repeated triggers belong in product or docs, not only in support.
Finding 04
Edge Case Exposure
—
—
Review repeat edge cases monthly. Three similar misses usually means a docs or product gap.
Finding 05
Out-of-Scope Rate
—
—
These are queries the bot correctly recognised as outside its scope. Real documentation gaps appear in Top Documentation Gaps below.
Finding 06
Top Documentation Gaps
—
—
These are your highest-value documentation gaps. Each one closes a blind spot the bot currently cannot handle — address the top cluster first.
—
—
—
Commercial Implications
Where the support signal points to margin, retention, risk, or growth
Highest-ROI Opportunity
—
of queries need a human today
—
—
—
current deflection
→
—
with automation fix
Sample Escalations
—
Sample Edge Cases
—
See how this was built, or explore a version for your business
ArNa Insights is a portfolio project built to show how I scope, build, and ship AI products that connect technical systems to business value. If you want to discuss what a version of this would look like for your support operations, I am happy to talk.
