Content teams don't usually stall on words — they stall on pictures. The article is written, the page is built, and then everything waits while someone hunts for an image that fits the brand, sorts out the licence, resizes it and uploads it to the right place. AgenticWebTeam folds that whole detour into the same chat where the content was made.
Three ways to get an image
The assistant has three image moves, and it plays them in a sensible order. First, reuse: your CMS asset library almost always contains more usable imagery than anyone remembers, and the assistant checks there before anything else. Second, stock: it searches professional photography and shows you candidates inline in the chat. Third, generation: when nothing exists, it creates an original image to your brief.
The order matters because it mirrors what a good editor does — cheapest, most on-brand option first. A photo you already commissioned beats a stock photo; a well-chosen stock photo often beats a generated one; and generation is there for the cases where only something bespoke will do.
Stock photography, without the tab-hopping
Ask for "a header image for the cadences article — something calm, office, not cheesy" and the assistant searches stock photography and presents a short shortlist inline, with the photographer's name attached to each. You pick one in plain language. The chosen photo is then imported directly into your CMS's asset library — correctly sized, named sensibly, with the photographer credit preserved — and dropped into the image field of the document you're working on.
Two details do a lot of quiet work here. Imports go into the same environment as the document that uses them, so a staging article gets a staging asset and promoting content never leaves broken image references. And the credit travels with the asset, so attribution survives however the image is reused later.
Generated images, on brief
For illustrations, abstract headers, product-adjacent artwork or anything photography can't provide, the assistant generates images from a description. The practical craft is in the brief: name the mood, the palette and what must not appear, the same way you'd brief a designer. Generated images land in your content library like any other asset, ready to place, and you can iterate — "same concept, less saturated, wider crop" — in the same conversation.
Keeping it on-brand
Brand consistency comes from three layers working together. Your assistant's instructions can carry standing art direction — palettes, tone, dos and don'ts — so every image request starts from your rules rather than a blank page. The conversation adds per-task direction. And because every image is placed as a draft change you review, nothing off-brand reaches the live site by accident: the same draft-and-publish ladder that protects your copy protects your imagery.
Permissions apply too. Placing an image is a content write like any other, so it's scoped by the requesting user's role and environment, and publishing it live requires publish permission — a junior writer can stage the perfect hero image without being able to ship it unreviewed.
A worked example
Here's the shape of a typical session. A marketing manager is finishing a case study and asks for "a header image — warehouse, morning light, people working, nothing staged-looking." The assistant checks the asset library first and surfaces two candidates from a shoot last year; neither fits. It searches stock photography, presents four options inline with photographer names, and the manager picks one with a one-line reply.
The import happens in the same breath: the photo lands in the CMS asset library at a sensible size, the credit is stored with it, and the case study's hero field is filled — as part of the same draft the copy lives in. When the manager later asks for a square crop variant for the social card, that's one more message, not a round trip through a design tool.
Total elapsed time is a couple of minutes, and the manager never left the conversation. Multiply that by every image on every page you ship in a quarter and the arithmetic gets persuasive quickly.
Rights, credits and good citizenship
Imagery comes with obligations, and the workflow bakes them in rather than leaving them to memory. Stock photographs carry their photographer attribution into your CMS as part of the import, so credit survives content restructures and team turnover. Generated images are original works created for your brief. And because every asset lives in your CMS rather than ours, your existing rights processes — licence records, usage reviews, takedowns — keep working unchanged.
Video, and what we don't do
A scope note, because vendors love to overclaim: this is an image workflow, not a full media production suite. There's no video editing and no pretence of replacing your design team for flagship creative. What it replaces is the coordination tax on everyday imagery — the ninety percent of images that need to be good, licensed and placed, not art-directed from scratch. Keeping that boundary honest is why the feature stays fast and dependable inside it.
The compound effect
None of this is spectacular on its own — search, import, generate, place. The compound effect is what changes the workflow: the image stops being a separate errand with its own tools, tabs and licence anxiety, and becomes one more sentence in the conversation that's already producing the page. Content that used to wait a day for artwork ships in the same session it was written.
Included skills are metered by the same credit system as everything else, so there's no separate subscription to manage — and if your organisation has its own stock or generation accounts, enterprise workspaces can bring their own keys and the assistant will use those instead.
