What Do AI Agents Publish at Work?
Display (display.dev) is our product. Display is the docs workspace for agents and people. Any agent publishes docs and files to a link. Your team opens the link with the account they already use. They comment and edit i
Display (display.dev) is our product. Display is the docs workspace for agents and people. Any agent publishes docs and files to a link. Your team opens the link with the account they already use. They comment and edit in-line. Any agent or teammate can publish the next version to the same link.
We give examples, not counts or other internal figures. We name two customers, Kogan.com and Indigo Engineering. All other examples are generic.
What the eight kinds have in common
Each kind of work has the same shape:
- A person asks an agent for a document.
- The agent publishes the document to a link.
- The team opens the link and comments on the part that they review.
- The agent reads the comments and publishes the next version to the same link.
The link does not change between versions. The comments stay on the page. Nobody reads an old copy.
1. Prototypes and design explorations
Product designers ask an agent for a clickable prototype or a design exploration. The agent publishes it. The team opens the link and comments on the screen that they review. For example, a comment can say "Put Skip for now under the main button." The agent reads the comment and publishes version 2.
Read Display for product designers.
2. Client reports and proposals
Agencies and consultancies publish reports, proposals and client work as a link instead of a PDF. The client opens the link with a one-time password that Display sends to their email. The agency team comments on the page. The agent publishes each revision to the same link.
An Owner can also restrict Public artifacts and outside sharing for the whole organization. This helps when a client contract limits where a report can go.
Read Display for agencies.
3. KPI reports and campaign plans
Marketing teams publish weekly KPI reports, campaign plans and landing page mockups. The team opens the link and comments on a chart or a slide. Next week, the agent publishes the new report as the next version. The link in the email from last week still works.
Read Display for marketing teams.
4. Architecture diagrams, feature prototypes and postmortems
Engineers ask their coding agent for a feature prototype, an architecture diagram or an incident postmortem. Product, design and support open the link and comment. They do not need a GitHub account.
Before Display, Indigo Engineering shared HTML artifacts in Slack. Max Wood, CEO of Indigo Engineering, describes the old way in one sentence. "For HTML artifacts, this meant a bunch of messy and often confusing screenshots." Read the Indigo Engineering story.
Read Display for software engineers.
5. Dashboards and analysis
Analysts publish dashboards and reports with working filters and charts. The business opens the link and needs no BI license. An agent runs on a schedule and publishes the next version, for example every Thursday. The link stays the same.
Kogan.com uses Display this way. Its IT team gave Display to a few people first. Then the rest of the business started to use it. Its dashboards refresh on a schedule at the same link. Rory Powell, IT and AI Ops Manager at Kogan.com, says: "We didn't realize the whole business would pick it up." Read the Kogan story.
6. Board updates and plans
Founders publish board updates, investor updates and plans. Only the people you add can open the link. The agent publishes the next quarter to the same link, so a board member always opens the current update.
7. Proposals for a prospect
Sales teams publish a proposal or a one-pager for a prospect. Only the email addresses you add can open the link. Each prospect signs in with a one-time password. When the offer changes, the agent publishes the new version to the same link. The prospect does not get a second attachment that conflicts with the first.
8. Meeting notes, plans and procedures
Operations teams publish meeting notes, offsite summaries and procedures. The whole company opens the link and comments. The agent publishes the corrected version to the same link.
What changes the result
Three choices are more important than the agent you use:
- Who can open the link. Choose public, everyone at your company, or the people you add. Readers use Google, Microsoft or a one-time password, depending on the plan.
- One link for all versions. Tell the agent to update the existing page. Do not ask it to publish a new page.
- Comments that the agent reads. The team comments on the page. Tell the agent to read the comments before it writes the next version.
What Display does not do
- Outside guests can view a document. They cannot comment yet.
- Folders, search and memory are on the roadmap. They are not available today.
- Custom domains and self-serve SAML are on the roadmap.
Try one
Pick the document that you send most often. Ask your agent to publish it to Display. Share the link with one teammate. Ask for one change in a comment. Then ask the agent to publish the next version.
FAQ
What can an AI agent publish at work?
An agent can publish any HTML or Markdown document: prototypes, reports, plans, dashboards, proposals, postmortems and meeting notes. Teams use Display for the eight kinds of work in this post.
Which agents can publish to Display?
Any agent that can write an HTML or Markdown file. This includes Cursor, Codex, Claude Code and local models.
Do readers need an account?
Invited people outside your organization do not. They open the link with a one-time password that Display sends to their email. Members of your organization sign in. On the Pro plan, they can sign in with Google or Microsoft.
Can I share a document with people outside my company?
Yes. Add their email addresses, or make the document public. An Owner can restrict Public artifacts and outside sharing for the whole organization.
Does the price change when more people read a document?
No. Every plan includes unlimited viewers. Read the pricing page.
Free tier. No credit card. One-time password auth for viewers on free, Google + Microsoft SSO on Pro (€49/month flat).