P1
Service Operating Model
LegacyAutonomous
- Legacy
- Agents handle every ticket by hand with no AI in the workflow.
- Autonomous
- Self-improving service flows resolve and escalate with minimal human staging.
AIR FOR CUSTOMER EXPERIENCE
Your Customers Will Meet AI Before They Meet Your Team
The AIR Index reads your support function across five pillars, so you can see where AI lifts service and where it quietly costs you trust.
AI now sits between you and every customer who reaches out, and the gap between teams that govern that layer and teams that do not is becoming the gap in retention itself.
What AIR measures
Five pillars, read for Customer Experience.
The same five pillars of AI readiness, framed in the work, systems, and stakes that CX and customer-support leaders actually face.
P1Operating Model
Whether AI is built into how the team actually handles contacts, with shared macros, summaries, and routing rather than tricks living in a few star agents.
P2Service Model & Outcomes
Whether AI changes what support delivers and how that value is measured, so deflection, resolution, and CSAT move together instead of trading off.
P3Talent & Capability
Whether agents and team leads can confidently edit, verify, and trust AI output in live conversations, and whether they see it as backup, not a threat.
P4Governance & Risk
Whether AI in service runs on real controls for transcript PII, answer accuracy, and bot disclosure, rather than hope that nothing wrong reaches a customer.
P5Tool Standardization
Whether the AI tools touching tickets, transcripts, and voice form a deliberate, owned, defensible stack instead of a sprawl no one fully maps.
The AIR rating
Six tiers, Legacy to Autonomous.
AIR places Customer Experience on a six-tier readiness ladder from 0 to 100, overall and for every pillar. The climb runs from digital but not intelligent, to a compounding, AI-native edge.
6
Autonomous
85–100AI-native advantage. Compounding intelligence and speed, a durable edge competitors can't copy fast.
5
Integrated
68–84Woven through the business. AI shapes the operating model, pricing, and talent, and ROI is proven.
4
Operational
51–67AI in the core, governed. Embedded at named steps with SOPs, policy, and measured gains.
3
Adopting
34–50Pockets, not a system. Real use in places, uneven and undocumented.
2
Reactive
17–33Experimenting at the edges. Scattered pilots that live in a few people's heads, ungoverned.
1
Legacy
0–16Digital, not intelligent. AI is absent or anecdotal, work is hour-priced, the stack sprawls, and no policy exists.
The point
One score tells you that you are behind; five tiers tell you exactly where to start.
The deep diagnostic
Every pillar, climbed for Customer Experience.
The matrix is a 5-by-6 grid: your five pillars of AI readiness scored against the same six tiers, from Legacy to Autonomous. A single overall score tells you roughly where you sit; it hides where you are dangerously behind and where you are quietly ahead. Reading a tier per pillar turns one vague number into five specific, fixable verdicts, so you act on the truth instead of an average.
P2
Service Model & Resolution Outcomes
LegacyAutonomous
- Legacy
- Output is human-written replies with no AI lift to speed or quality.
- Autonomous
- The function predicts, prevents, and resolves, turning service into durable advantage.
P3
Talent & AI Capability
LegacyAutonomous
- Legacy
- No agent is expected or trained to use AI in their work.
- Autonomous
- The team continuously builds AI capability as a compounding service asset.
P4
Governance & Customer Risk
LegacyAutonomous
- Legacy
- No policy governs AI in customer contact and risk is unmanaged.
- Autonomous
- Trustworthy automation is a competitive asset with provable, audited safety.
P5
Tool Standardization
LegacyAutonomous
- Legacy
- No AI tools are licensed for support beyond the core helpdesk.
- Autonomous
- A defensible, proprietary stack and data flywheel widen the service lead.
Governance and compliance
Where the rules bite.
How Customer Experience's regulatory reality maps onto AIR readiness. Each row is a control your governance pillar has to carry.
| Regime | Pillar | What AI readiness requires |
|---|---|---|
| FTC Act / consumer-protection norms | P4 | AI-handled service interactions must not be unfair or deceptive, and AI-generated promises bind the company. |
| State chatbot-disclosure laws (e.g. California B&P 17941 "bot" rule) | P4 | Disclose to customers when they are interacting with an automated agent rather than a human. |
| GDPR / CCPA-CPRA (PII in transcripts and voice) | P4 | Lawful basis, retention limits, and deletion rights for call recordings, chat logs, and PII fed to AI tools. |
| PCI DSS (payment data in support channels) | P4 | Keep cardholder data out of AI prompts, transcripts, and tool logs in payment-handling support flows. |
| ADA / WCAG 2.2 (accessible service channels) | P4 | AI chat, voice, and self-service experiences must meet accessibility standards for all customers. |
| TCPA / call-recording consent (two-party states) | P4 | Capture consent before AI transcribes, analyzes, or synthesizes recorded voice interactions. |
Illustrative mapping for AI-readiness planning, not legal or compliance advice; validate against current regulation with qualified counsel.
The stakes
What stalling looks like.
Confidently wrong answers reach customers
An AI agent invents a policy, refund amount, or product detail and a customer acts on it. Without a QC checkpoint, the hallucination becomes a commitment the company has to honor or publicly retract.
PII leaks through transcripts and prompts
Agents paste full chat logs, call recordings, or account details into consumer AI tools to summarize them. Customer names, payment data, and health or financial details land in ungoverned systems with no retention control.
Tone and empathy collapse at scale
AI-drafted replies sound generic, miss the emotional read of an upset customer, or apply the wrong brand voice. Deflection rises while CSAT and trust quietly erode underneath the metric.
Undisclosed bots erode trust and break rules
Customers cannot tell they are talking to AI, or discover it mid-conversation. Beyond the trust hit, several jurisdictions now require disclosure that an interaction is automated.
Deflection masks unsolved problems
Containment and ticket-volume metrics improve because the AI closes conversations, not because issues are resolved. Repeat contacts, churn, and escalations rise while the dashboard looks healthier than the customer experience.
Start now
Signature quick wins for Customer Experience.
Concrete first moves you can make before the full diagnostic, one per pillar where it matters most.
P1Weeks
Turn your top intents into agent-assist macros
Take the 10 highest-volume contact reasons and build shared, reviewed AI reply drafts and summaries into the helpdesk, so quality stops depending on which agent is online.
P4Days
Ship a one-page CX AI policy
Write what AI tools agents may use, what customer data never gets pasted, and when AI use must be disclosed, then have the whole team read it this week.
P2Weeks
Define a deflection-with-resolution metric
Pair every containment number with a downstream repeat-contact and CSAT read, so the function reports problems solved, not just conversations closed.
P5Days
Audit the support AI stack in 30 minutes
List every AI tool touching tickets, transcripts, or voice, name an owner and the data each one sees, and kill the most obvious redundancy.
P4Days
Add a human QC gate on AI answers
Require that AI-drafted responses on refunds, policy, and complaints get a human check before sending, to catch confidently wrong or off-tone replies.
P3Days
Run a role-specific agent-assist session
Hold one 60-minute working session where frontline agents learn to edit, verify, and trust AI drafts, reframing the tooling as backup not replacement.
Find out where your team stands.
The old transformation is finished. The new one is scored.