Making changes in Optiply for you
The agent can do in Optiply what you can do in the app, using your own user permissions:
- Suppliers: settings such as lead time, reorder period, order rules, minimum order value.
- Products and supplier products: settings, links between products and suppliers, predecessor links (an old product linked to its replacement), purchase prices, MOQ and lot sizes.
- Draft (concept) orders: save a whole proposed order as a draft, or build a draft for one supplier from specific SKUs and quantities (up to 200 lines per draft).
- Order lines of drafts and proposed orders: change quantities, add products or remove lines on an existing draft order, or on a proposed order before it's placed (up to 200 line changes per approval). Changes to a proposed order are saved as a new draft order, just like Save in the order screen; placing that order is a separate step that you approve on its own.
- Purchase orders: place orders, update order lines, close orders, export settings.
- Deliveries: register (partial) deliveries.
- Order moments: plan and replan the order agenda of a supplier.
- Promotions: create and update promotions and link or unlink products.
- Webshop settings.
- Service-level simulations (these change nothing).
A few settings screens of the Optiply app can only be changed in the app itself. For those, the agent tells you where to make the change instead of proposing it.
How changes stay safe
- You approve every change. The agent shows a plain-language summary, for example *"Change the lead time of supplier Acme BV to 7 days"*, with the current value read live from Optiply next to the new one, and waits for you to click Yes or No. In the app the proposal is a card marked Waiting for your approval, which afterwards shows Approved or Not now. A change to many records lists every record by name; a new promotion shows whether it is active right away and what it applies to. Nothing is changed before you click Yes. If you click No, it won't propose the same change again. If you reply with a change instead ("make it 50%"), nothing is applied yet: the agent proposes the whole set again with your change.
- It checks before asking. Required fields, allowed values and whether a product can actually be ordered from that supplier are checked before you see the proposal.
- It warns you about lasting effects. For example, that a supplier setting affects all future advice, or how many lines of a purchase order are still undelivered before you close it.
- It checks the result. After a change, the agent reads the record back from Optiply and only tells you it's done once it has confirmed the new value. If it can't confirm, it tells you so and asks you to check in Optiply. When the agent proposes a next change in the same reply, the proposal first lists what happened to each change you just approved (done, not done, or please check). If Optiply did not accept a change, the agent tells you instead of quietly trying another way; ask it if you want it to try differently.
- One approval for repeated changes. When the same change applies to several records, you approve them together, and the agent then makes them one by one.
- Bulk changes: Imports, or rounds of 50. For more than 50 records (suppliers, supplier products, products, promotions, compositions or sales data), the agent recommends Settings › Data Management › Imports/Exports: it handles everything in one go, and the agent can prepare a pre-filled CSV for you. After the import, ask the agent whether it went through: it reads the result of your recent imports, with the rows imported, failed and skipped and the reason for each skipped row. If you'd rather do it in the chat, say so and the agent makes the change in rounds of up to 50 records, with one confirmation per round. Changes the import can't do, such as linking many predecessors, go through the chat the same way: paste your list or upload your Excel or CSV (for predecessors: old SKU and new SKU) and the agent looks up every product and shows you all the links to approve.
Asking *how* to change something works too: when the agent can make the change itself, it offers to do it for you first, and then explains where to do it in Optiply if you prefer that. If you say you want to do it yourself, it just gives you the steps. With read-only access to the agent, it explains the steps instead.
Read an approval card
Where a proposal card is offered, review its changes and choose its approval or cancellation button. The exact labels come from that proposal; if no card is shown, use the current answer's offered choices instead. While your reply is being answered, the card can show Saving. Older unanswered cards and cards whose follow-up failed are read-only: do not treat them as a new approval opportunity.
Approved or Not now describes your reply on the card, not independent confirmation that every requested change was saved. Read the follow-up and check the affected record in Optiply. Likewise, Done on a chat tab or work-step summary is not a successful-change receipt. If the reply failed or the outcome is unclear, check the current record before approving or retrying another request.
What it won't change
- Values Optiply calculates or measures itself: measured lead time, stock on hand, safety stock.
- Empty values that Optiply always needs, such as a supplier's lead time. You can change them to a new value, but not leave them blank; the agent explains this instead of proposing the change.
- Data that comes from your integration (your ERP, WMS or webshop). Changing it in Optiply would be overwritten by the next sync, so the agent explains the difference and, where needed, creates a support ticket instead.
- Anything your user role isn't allowed to change.
Explore the Optiply Agent guides
- Get started with the Optiply Assistant
- Analyze purchasing, stock and suppliers with the Optiply Agent
- Work through plans with the Optiply Agent
- Use files, exports and charts with the Optiply Agent
- Schedule jobs and email reports with the Optiply Agent
- Optiply Agent memory, privacy, permissions and limits
- Connect the Optiply Agent: MCP and Slack setup