LoginSign Up
← Back to all sparks
A

Atarim

PM
Velocity5.0

Visual collaboration and work management platform for web agencies, enabling point-and-click feedback, task management, and client approvals on live websites.

atarim.io ↗

Atarim's AI experts now propose changes to live pages as drafts, and the app runs on phones.

agency-workflowai-agentshuman-in-the-loopmobilewordpress
◆Current state
Atarim 5.1.4 extends the draft-first rule from new pages to edits on live pages: an AI specialist's fix comes back as a draft copy with a preview link, and the task waits in Pending review until a person approves it. Atarim now also works on phones and tablets and installs as an app. The previous release, 5.1.3, rebuilt Boards with a timer on every stage.
◆Where it's heading
Atarim keeps pairing more AI autonomy with more human control: readable memory in 5.1.2, stage clocks in 5.1.3, and now mandatory approval on live changes. The mobile app moves that approval step to wherever the agency owner is, which fits a model where AI does the work and people sign off.
◆Prediction
Expect approval and review to tie into Boards, for example pending AI drafts showing up as a stage with its own clock.

◆Recent moves

  1. 6d ago

    Atarim 5.1.4: your AI team proposes, you approve, from anywhere

    Changes to live pages now arrive as drafts with a preview link and wait for human approval, extending 5.1.3's draft-first rule for new pages. Mobile and installable app support lets agencies approve from anywhere.

  2. 22d ago

    Atarim 5.1.3: the new Boards are here, with a clock on every stage

    This is a re-publication of the 5.1.3 Boards release dated a day later; the content is the same. Boards with per-stage timers give agencies a view of where client work has stalled.

  3. 23d ago

    Atarim 5.1.3: the new Boards are here, with a clock on every stage

    5.1.3 ships the rebuilt Boards: agencies define the stages their client work moves through and every project, page, or task shows how long it has sat in its current stage. It applies the V5 accountability theme to human workflow, not just the AI team.

  4. 1mo ago

    Atarim 5.1.2: your experts' memory is now yours to read and edit

    ⚡ SPARK

    The memory the AI experts have built since 5.1.0 is now readable and editable by the agency. It is the strongest move yet in Atarim's push to make its agents auditable rather than opaque.

  5. 1mo ago

    Nothing says "done" unless it is done (5.1.1 and plugin 5.1.3)

    Atarim hunted down four places where the product reported success without doing the work, from workflow steps to skipped monthly jobs and backups. It's a reliability release, but it directly supports the claim that the AI team's output can be trusted.

  6. 1mo ago

    Cheaper AI, real SEO research, and work you can actually check (5.1.0)

    This repeats the 5.1 notes: about 40% cheaper AI actions, no charge for failed turns, SEO research, and reports that log their sources. The substance was covered by the original 5.1 release; this entry adds no new capability.