Use AI autofill for the judgement calls an agency makes on every request, formulas for the rules, and keep what your team typed. Setup, cost, examples.
Attach the client's brief PDF to an AI agent, get the deliverables, deadline, approver and open questions back, then check each against the line it came from.
An AI agent skill is a reusable set of instructions and tools for one kind of work. Write your agency's process once and share it across agents.
Before an AI agent updates more than 20 rows by filter, you see a preview; bulk changes keep a 7-day Undo; emails and Slack posts wait for your approval.
AI autofill, an email address for every sheet, guest support tickets, agent skills and agents that read files: what changed in Atwork in October 2026.
Dana is the AI assistant built into Atwork. It answers questions about your client work, builds sheets, writes pages and plans your day, as you, on your data.
An Agent step reads what a rule cannot, like a client email or a brief, and turns it into fields and branches the rest of the workflow acts on.
Put a researcher, a writer and an editor agent in one workflow, see what each step costs, and keep a person between the draft and LinkedIn.
An AI-drafted client update can remember what you promised last week. What it keeps, for how long, where to see it, and how to forget one client.
An approval belongs where an AI agent decides what to change. Here is where to put one in a workflow, who gets it, and what happens if nobody answers.
An Agent step reads each new client request and labels it in scope, change request or new project, so a change gets priced before anyone does the work.
What can genuinely run unattended in agency client reporting, what still needs a person, and the one status setting that decides whether it works.
Which marketing-agency jobs you can hand to an AI agent today, which you cannot, and where the line is — including what it can and cannot publish to.
An AI agent does not fix an operations problem — it compounds whatever your operations already are. What to fix first, and how to tell whether it is working.
Four ways to start an AI agent — assign the task, mention it in a comment, move the status — plus how to stop it and what a workflow can hand over.
Vista is an Atwork view where you describe a page and Dana builds it on your live sheets: dashboards, forms, and client pages you can publish as apps.
Assign a task to an AI teammate the way you would a person, then read the trace: every step, how long it took, and which actions needed your approval.
Connect your Atwork workspace to ChatGPT or Claude over MCP and ask about your real client work. It acts as you; sensitive actions are not offered.
Per-seat AI uplift charges for people; an agency's unit is work. How metered per-run cost, your own provider key, and per-agent caps change the maths.
Ask your AI assistant about your Atwork tasks, sheets and dashboards without opening Atwork. What it can do, what it cannot, and how permissions stay safe.
Build a custom AI agent in Atwork, scope what it can see, assign it tasks like a colleague, and @mention it in comments. A practical setup guide.
Five things every AI agent persona should spell out, with before-and-after examples you can copy. Written for people configuring agents, not engineers.
Scope, capabilities, and identity: how to reason about what an AI agent can read and change in your workspace before you roll one out to your team.
Atwork speaks MCP, so Claude can read your sheets and create tasks directly. What it can do, what it cannot, and when this beats using an in-app agent.
A practical rule for deciding between a workflow automation and an AI agent, plus a comparison table covering cost, reliability and how each one fails.
Step-by-step: scope, persona, statuses and review loop for an AI agent that triages incoming support requests and hands each one back to a human.
A practical way to pick the model behind your AI agent: match the tier to the job shape, start cheap, and upgrade only when you can point at a failure.
Running an MCP server in production is mostly authorization and tool design, not protocol work. Lessons from putting a real workspace behind MCP.
5 Must-Have Productivity Tools for Startups: Why Atwork Stands Out Launching and scaling a startup is an exhilarating journey, but it comes […]
In today’s fast-paced world, managing tasks efficiently is the key to success. Whether you’re a freelancer, a startup founder, or part of […]
If you’re the one responsible for keeping track of company meetings, here’s everything you need to know about writing effective work ad.
This is another way of saying, “We’ve been in business for a long time because people love our products and trust us.”
Again, customers liked to be affirmed in their decisions. This phrase communicates that if the customer buys your products.