It can read the ticket, check the affected service in the browser, gather the relevant policy and audit evidence, and prepare the next step for review. It should not quietly grant access, remove an account, change a configuration, or reset a credential without a named approval rule.
01
Start with the evidence, not the button.
A request to add a user, investigate a login failure, or retire a SaaS seat often arrives with missing context. The useful first pass is to collect the requester, system, role, manager, policy, prior tickets, and exact scope before anyone opens an admin control.
02
Incidents need a timeline before they need a summary.
During an outage or security concern, scattered tabs are often the actual record: a status page, a ticket thread, a monitoring dashboard, a vendor notice, and a chat channel. Strawberry can assemble those sources into a time-stamped brief that distinguishes observed facts from unanswered questions.
03
SaaS hygiene is mostly a matching problem.
License reviews fail when names, email addresses, HR records, and vendor dashboards disagree.
Ask for stable identifiers, a defined inactive threshold, and an exceptions list instead of letting an agent infer who should lose access.
04
The handoff should be more useful than a ticket update.
An administrator often needs to leave a decision record: what was checked, what changed, why it was allowed, and how to reverse it. Strawberry can prepare that record in the systems your team already uses. If Slack is connected, its supported operations include searching messages, reading a channel or thread, and sending a message after review.
AI can investigate requests, gather evidence, prepare documentation, and stage administrative work. Keep consequential system changes behind an administrator’s approval.
It can work in the browser session where you are already signed in, subject to the controls and permissions of that environment.
Yes. It can compare exports and approved employee data, then separate matches from duplicate or uncertain identities for human review.
With Slack connected, supported operations include searching messages and reading channels or threads, which can help assemble incident context.
You can save a defined review as a skill and schedule it as a routine, provided the data sources, scope, and escalation rules are clear.
No. Use it to surface inactive or unmatched accounts, then require a responsible administrator to review any removal.