MorningAI agents start every conversation already knowing your brand. There's no pasting of guidelines, no re-explaining your products, no describing your audience from scratch — the grounding is built into the product and applied automatically. This article explains every context source the agent draws on, and what it remembers as a conversation goes.
The Six Context Sources
| Source | How it works for you |
|---|---|
| BrandDNA | Applied automatically to every conversation — voice, style, brand facts. Nothing to attach. |
| ProductDNA | @-mention a product to attach it to a turn; the agent can also search the catalog itself, down to variants and images. |
| CustomerDNA | @-mention a persona as the audience for a task, or speak to a persona directly via Customer Insights mode. |
| Knowledge Base | Documents uploaded to your brand's Knowledge Base are retrieved automatically when they're relevant to the question. |
| Attached files | Drop files into the composer; the agent reads them and can re-read them later in the conversation. |
| Conversation memory | The agent remembers the conversation and everything it made — data pulls, images, charts, files — and can find and reuse them in follow-ups. |
Mentioning: @ Products and Personas
Type @ in the composer (on Home or in Chat) and a dropdown opens with tabs for your products and personas. Mentioning a product attaches its full profile — details, DNA, images — to that turn. Mentioning a persona sets the audience: "write launch copy for @Budget-Friendly Brenda" aims the work at that segment.
The distinction worth remembering: @-mentioning a persona makes them the audience of a task, while Customer Insights mode makes a persona the speaker — the agent answers as that customer so you can test messaging.
Memory Within a Conversation
Everything the agent produces in a conversation stays available to it. "Use the chart from earlier in the deck" works. "Rewrite the summary from that file you made" works. After a turn, useful content is saved — you'll see "Saving to memory" in the progress chip, then "Saved to memory" when it completes.
Set Up the Grounding Once
The richer your DNA, the better every agent answer gets. Each source has its own setup guide:
- Set up BrandDNA — your voice, style, and brand facts.
- Build your ProductDNA catalog — products, variants, and images.
- Set up CustomerDNA — your researched customer personas.
- Fill your Knowledge Base — guidelines, research, and reference documents.
Frequently Asked Questions
Does the agent's work reflect my playbooks and personas too? Yes — playbook instructions and attached personas shape generated files and images the same way they shape chat replies.
Can the agent create personas or products from chat? Yes: it can create new personas, and create or update products (including uploading images). It can never delete either.
Next Steps
Try one grounded request — "@-mention" your best-selling product and ask for three campaign angles — then explore the full capability map in Meet MorningAI Agents.
