Grow cluster · Support, customer success
Technology Rewired: Customer Support and Success
Customer support is the category where AI agents are furthest along. Software moved from on-prem ticket queues to SaaS help desks, then to bots that deflected questions. Today agents resolve routine tickets end to end, refunds and resets included, while human teams shrink toward exceptions, escalations, and the customers worth keeping.
The shift: Tier-1 resolution moves from humans to agents; humans handle exceptions.
Agents with approvals today
4.0 in five years
How has the customer support and success team changed across five eras?
Era 1 · On-prem
Before 2005
0.2Support was a phone room and a shared inbox. Tier-1 agents answered calls from scripts, tier-2 handled anything technical, and a supervisor watched queue length on a wall board. Ticketing ran on server software the IT team maintained, and headcount scaled almost linearly with customer count.
Era 2 · SaaS and cloud
2005 to 2020
0.8Browser help desks put email, chat, and phone in one queue and made knowledge bases public. Teams split into tier-1 generalists, specialists, a knowledge manager, and the first workforce planners. Customer success appeared as its own function for SaaS renewals. Pricing was per seat, so every new agent hire was also a new license.
The pyramid stayed the same shape: a wide base of people answering the same questions every day.
Era 3 · AI-assisted
2020 to 2024
1.6Chatbots and suggested replies arrived. Most bots deflected rather than resolved: they pointed customers to articles and handed off when the question got real. Agents got macros, summaries, and sentiment flags, but a person still closed every ticket that mattered.
Era 4 · Agentic
2024 onward
3.2Agents now resolve password resets, order changes, refunds within policy, and account questions end to end, inside the help desk and the systems behind it. Teams are reorganizing around the exceptions: senior agents who take escalations, a conversation designer who owns agent behavior, and an operations lead who watches resolution quality. The per-seat model breaks when the best worker is not a seat, which is the argument I make in AI killed the SaaS seat.
The new risk sits at the handoff. A support agent with refund and reset powers is a target, which is why the help desk is the new perimeter.
Era 5 · Next 5 years
2026 to 2031
4.0My bet: the default support team becomes a small group of specialists supervising agents that handle the majority of volume across chat, email, and voice. Success teams spend their time on accounts that are growing or at risk, using agent-written account briefs. Headcount stops tracking customer count.
Which customer support and success software is being rewired?
- Help Desk and Ticketing
Help desk software is moving from a queue that routes tickets to people into a system where AI agents resolve most routine requests end to end. The bigger shift is pricing: vendors like Intercom and Zendesk now bill per resolution instead of per seat. Humans keep exceptions, judgment calls, and anything touching identity.
Coming next
- Live chat and conversational supportwave 2
- Knowledge base and documentationwave 2
- Chatbots and voice agents
- Call center and CCaaS
- Community forums
- Customer success platforms
- NPS and surveys
- QA and conversation analytics
- Field service management
Also wired into this category: Incident Management and On-Call, Customer Onboarding Email.
Keep reading
Essays and analysis
- AI Didn't Kill SaaS. It Killed the Seat.
- Your Help Desk Is the New Perimeter: What MGM, M&S, and Scattered Spider Taught Every IT Team
- What AI Actually Changes About the Sales to Support Handoff
- Adobe's BPO Breach: One Phished Support Agent Extracted 13 Million Records and Unpublished Vulnerability Reports
- The Post-Agentic Organization: What Becomes Scarce When Intelligence Is Abundant