A CS agent reads three surfaces on a new account. Tickets in Zendesk. Health and adoption in Gainsight. Deal context in HubSpot. None of them holds the agent's interpretation of the hand-off. Dock does. The agent drafts a brief in a Dock row, the CSM accepts or rewrites it, and the audit attaches to the agent identity. See the pillar at Dock for customer support.
Zendesk, Gainsight, and HubSpot stay the system of record for the raw data. Dock is the system of record for what the AGENT INTERPRETS. Each Dock row carries a pointer back to the platform record, agent identity, decision, reviewer, and timestamp. The agent re-fetches platform data via fresh API reads when it needs current state.
The Dock surface: onboarding hand-off table
| account | zendesk_signals | gainsight_health | hubspot_deal | agent_draft_brief | csm_reviewer | status |
|---|---|---|---|---|---|---|
| Lumen Logistics | 4 setup tickets, 2 closed | Yellow, adoption 38% | Mid-market, 14-seat | Risk: SSO blocker. Push kickoff to week 3. Owner: Priya. | @rachel.cs | accepted |
| Northwind Foods | 1 ticket, billing q | Green, adoption 71% | SMB, 6-seat | On track. Schedule QBR at day 60. Light-touch pod. | @marco.cs | rewritten |
| Habitat AI | 7 tickets, 3 open | Red, adoption 12% | Enterprise, 80-seat | Escalate. Exec sponsor sync needed in 48h. | @rachel.cs | accepted |
Each row is the agent's reading. The platforms still hold the truth.
The workflow
The agent runs each weekday morning. It pulls open Zendesk tickets for accounts inside their first 90 days, the current Gainsight health and adoption score, and the HubSpot deal notes. It writes a draft brief into the Dock row: what stage onboarding is at, what is blocking, which CSM pod fits, what the first CSM action should be.
The CSM opens Dock, sees the queue filtered to their pod, and reviews each draft. They accept, edit, or rewrite. Dock keeps both versions: the agent's original and the human's final. The accepted brief writes back to Gainsight as a timeline entry tagged with the agent identity. The CSM owner writes to HubSpot. Zendesk gets a macro reply only if the CSM checks that box.
Nothing fires without a reviewer. The agent does not close tickets, change health scores, or send customer email. It interprets and proposes. See how to run customer support with AI for the broader pattern.
Why it matters
Onboarding hand-off is where CS quality leaks. The reading of what an account needs gets reconstructed from scratch each time. An agent can carry that reading forward, but only if its draft is reviewable and attributable. Gainsight's Customer Success Index 2025 reports 52 percent of CS teams now use AI in workflows and 91 percent expect moderate or significant impact on strategy. ChurnZero's Customer Revenue Leadership Study found teams with integrated CSM, support, and account-management coverage report higher NRR. AI proposals need a reviewer seat, and that seat needs a surface that is not the source platform.
When a brief is wrong, the team needs to know which agent drafted it, against which platform data, at what time. That is the agent audit record. The name on the row is the agent identity, not a service account borrowed from a CSM.
Start a Dock workspace for your CS team.
FAQ
Does the agent change Zendesk, Gainsight, or HubSpot directly? No. The agent reads from those platforms and writes its interpretation into Dock. Only after a CSM accepts a brief does Dock write back to the source platform.
What if the CSM disagrees with the draft? The CSM rewrites the brief in Dock. Both versions stay on the row, which makes the audit log useful months later.
Can a marketing agent share the same Dock workspace? Yes. The pattern mirrors lifecycle hand-offs from marketing to CS. See Dock for marketing.
How does the agent prove it read fresh data? Each row stores the API read timestamp for Zendesk, Gainsight, and HubSpot at draft time. The agent re-fetches on review if the row sits longer than four hours.