Support is the operational cost with the shortest AI payback. AI-resolved tickets cost $0.62 vs. $7.40 for human agents. Portfolio companies deploying tier-1 automation cut blended cost-per-ticket by 60–80% inside a quarter — while raising CSAT on the intents AI handles best. This is the play that shows up on the P&L in the first quarter after go-live.
Every additional $10M of ARR at most portfolio companies adds 3–5 support headcount. The unit economics get worse, not better. Meanwhile a huge share of tickets are structurally identical — password resets, order status, refund lookups — and cost $7 per resolution to handle manually.
We deploy inside Intercom, Zendesk, Freshdesk, HelpScout, or GHL — whatever the portfolio company already uses. AI handles the structured intents where it wins on speed and CSAT. Sentiment-heavy or complex tickets route to humans with full context attached. Nobody logs in to a new tool.
Blended cost per ticket falls from ~$12 to under $3 in the first quarter. The support team is redeployed to tier-2, customer success, and retention — the motions that actually drive expansion revenue. The exit story: automated support with a customer success bench, not a growing cost center.
Ranges reflect published 2026 benchmarks (McKinsey AI in Customer Service, Fin.ai ROI Benchmarks, Lorikeet CX Metrics) and typical Alterra AI portfolio deployments in SaaS, DTC, and services.
Sources: McKinsey AI in Customer Service 2026 ($0.62 AI vs. $7.40 human per resolution), Fin.ai 2026 ROI benchmarks (340% first-year ROI median), Lorikeet CX Metrics 2026 (41.2% median tier-1 deflection, 58.7% top quartile). Portfolio-wide rollouts realize an additional 40% cost efficiency vs. one-off deployments.
Every inbound ticket classified in real time by intent, sentiment, and complexity. AI-resolvable intents go to AI. Sentiment-heavy or complex tickets route to humans with the classification, context, and suggested response already attached. Nobody sees "which queue is this for?" again.
Chat and email tickets resolved end-to-end for the structured intents where AI outperforms humans on speed and CSAT — password reset, order status, refund lookup, shipping updates, billing questions. Runs 24/7 without a weekend premium.
Voice AI handles inbound support calls for the same structured intent set. Full conversation transcript logged to the ticket. Handoff to human agent with context if the call requires escalation. Cuts avg call handle time and eliminates queue times for tier-1 volume.
Every resolved ticket becomes a candidate KB article. AI drafts the article from the resolution transcript, human approves. Existing KB continuously refreshed based on what tickets keep coming in — so the deflection rate keeps climbing quarter over quarter.
48-hour deep dive on 90 days of ticket history.
Deploy inside existing help desk (Intercom, Zendesk, etc).
Watch live traffic. Widen the AI-resolution scope.
Baseline vs. current. Reallocate team.
Ranges reflect typical outcomes for portfolio companies handling 5,000–50,000 monthly tickets with a functioning-but-manual support motion. Portfolio-wide rollouts standardize these KPIs across every portfolio company for fund-level dashboards.
Get the assessment →The median tier-1 deflection rate is 41.2% in 2026, top quartile 58.7%. Structured intents (password reset, refund status) deflect at 70%+. Sentiment-heavy tickets (complaints, disputes) rarely exceed 25%. Target a 45–55% blended deflection rate in the first 90 days, focused on the intents where AI-resolution CSAT is highest.
AI-resolved tickets average $0.62 vs. $7.40 for humans (McKinsey 2026). Chat AI runs ~$0.41, voice AI ~$1.18. Blended cost per ticket at portfolio companies typically falls from ~$12 to under $3 in Q1. First-year ROI averages 340%.
Pure-AI handling is 4.1/5 CSAT vs. 4.3 for humans. Hybrid flows — AI on structured intents, human handoff for sentiment-heavy — narrow the gap to 0.05 and often outperform humans on speed. The failure mode is deploying AI on tickets it can't resolve. Intent-level routing and CSAT-triggered handoff are non-negotiable in our builds.
Best-in-class outcome is reallocation to tier-2 escalations, customer success, and retention motions. A 20-person team typically becomes a 6–8 person tier-2/success team plus AI, running 3x volume with better outcomes. Headcount reduction is possible but rarely the highest-ROI move — freed capacity applied to retention/expansion returns more than salary savings.
Resolution. Deflection means the customer gave up. Resolution means the issue actually closed without a human. Vendors quoting deflection numbers without CSAT tend to overstate savings by 20–30 points. We build for measured resolution + CSAT, not deflection.
48-hour AI opportunity assessment. Ranked by EBITDA impact. Delivered before the next partner meeting.
Book an assessment →Response within 1 business day. Fixed-scope, fixed-fee engagements.