Databox's new Routines feature schedules a saved analysis to run and deliver itself, but the capability only unlocks on a $199-a-month annual plan.
A marketer running a client account or a growth channel usually owes someone a report every week. Pull the ad spend from one dashboard, the conversions from another, the web traffic from a third, then stitch it into slides and flag whatever moved. It is the kind of work that has to happen on a schedule, and it rarely feels like the best use of a skilled person's Monday. A Google Cloud customer story about Supermetrics puts a number on that cycle: performance marketers there were spending up to four hours a week manually pulling data, formatting slides, and hunting for the cause of a performance dip, before the company automated its own internal reporting and said it reclaimed more than fifteen hours a month per marketer. That is one company's account of its own workflow, not a universal figure, but it names the kind of hours this category of work tends to eat.
Databox, an existing business intelligence platform, just extended its AI Analyst product with a feature called Routines. The pitch, laid out on Databox's agentic analytics page, is straightforward: you save how you already run an analysis as a "Skill," reusable steps and standards that anyone on the team can trigger. A Routine then puts that Skill on a schedule, daily, weekly, monthly, or a custom cadence, and delivers the finished report to email, Slack, or in-app before anyone has to ask for it. Databox frames this as the second layer of a three-layer stack: Skills capture how the work gets done, Routines run it on a timer, and a separate "Agents" layer, still marked "coming soon" on the same page, is meant to eventually carry a workflow from analysis into action. Routines itself is the part that ships now.
The pull of this is obvious for anyone who has ever built the same report by hand more than twice. The catch is where the feature sits on Databox's pricing page. The Free plan gives one user three data sources and 50 AI credits a month, no Routines. The Analyst plan, $71 a month billed annually, adds AI credits and custom integrations for a solo user, still no Routines. Routines only appears once you reach the Team plan's Core tier, $199 a month billed annually, which covers three users, ten data sources, and 500 AI credits a month. A larger Scale tier at $319 a month billed annually raises that to ten users, thirty data sources, and 1,000 AI credits. Extra data sources beyond a plan's included count cost $5.60 a source per month on annual billing. None of Databox's own materials attach a measured time-savings figure to Routines specifically, so the fifteen-hours-a-month number above is a benchmark for the category of work, not a claim Databox is making about its own customers.
That price and structure point to who this is not built for. A solo consultant or a single in-house analyst on the Free or Analyst plan gets the AI Analyst chat feature but not the scheduling layer, so the manual weekly build still has to happen by hand unless the budget stretches to $199 a month. An agency juggling a long roster of small clients will also want to look closely at the per-source add-on cost and the separate Agency pricing track before assuming Team is the right fit. And any team that needs the report to trigger a downstream action, not just land in an inbox, is waiting on the Agents layer that Databox has not shipped yet.
What Routines actually buys, today, is the difference between a report someone has to remember to build and one that shows up whether or not anyone remembered. That is a smaller claim than "AI runs your marketing analytics," and it is the more honest one. The open question for any team considering it is whether the report itself is still the right unit of work, or whether the real cost was never the pulling and formatting, but the closer look at the numbers that a rushed weekly deadline never left room for.