Chat
Chat is the command centre of dpaperwork. Instead of clicking through menus to find information or trigger actions, you ask — in plain language — and dpaperwork responds with live data, takes action, or both.
What you can do in Chat
Query your Workspace
Ask anything about your data and get a structured answer:
- "How's my pipeline looking this week?"
- "Which clients haven't heard from us in 14 days?"
- "What tasks are overdue on the Acme project?"
dpaperwork reads across all your Domains and renders the answer as text, a table, or inline cards — whatever fits best.
Create records
Describe what you want to create and dpaperwork does it:
- "Add a new lead — Bright Agency, contact is Sarah Chen, sarah@brightagency.com"
- "Create a project for Nova Corp — kick off 15 July, deadline 30 September"
Trigger Playbooks
Run any of your saved Playbooks on demand:
- "Run the weekly client report Playbook for all active retainers"
- "Send the onboarding checklist to Bright Agency"
Draft documents
Ask dpaperwork to draft a client-facing document using the relevant context:
- "Draft a proposal for the Nova Corp deal — $8k, 3-month engagement"
The draft pulls in your Company Context, the deal details, and your Brand Voice automatically.
Log decisions
When you make a strategic call in conversation, dpaperwork recognises it and offers to save it to your Company Knowledge:
- "We're going retainer-only from next month"
- "Focus on healthcare clients this quarter"
- "No more fixed-price projects under £5k"
Once saved, these decisions are recalled automatically in future sessions — you never have to repeat context you've already given.
Company Knowledge recall
Before every reply, dpaperwork searches your Company Knowledge for relevant context: past activity (closed deals, sent documents, completed projects) and saved decisions. This is what makes Chat feel like talking to someone who actually knows your business.
You can also query your memory directly:
- "What did we decide about pricing?"
- "What's the history with Acme Corp?"
Choosing a model
Above the composer, a model switcher lets you choose which AI powers the conversation — Deepseek, OpenAI, or Google models. The choice applies across the Workspace and you can change it any time. Different models are good at different things: the default is a fast, capable balance; the larger models are worth reaching for on long documents or tricky reasoning.
Working with your team
Conversations aren't private by default — you can turn a one-person exchange into a shared conversation with your team.
Making a conversation shared
Click Invite in the conversation header, pick a member, and they're in. Everyone who joins can read the whole thread and reply — like a group chat that dpaperwork is a part of.
- The header shows a participant count and an avatar stack — click an avatar to see who's who, and to remove someone or leave the conversation yourself
- Presence — you can see when other people are viewing the conversation and typing
- Reactions — hover a message and hit the quick-react bar to add an emoji; reactions sync live across everyone
- Mentions — type
@Nameto direct a message at someone (they get notified), or@dpaperworkto make sure the assistant replies - dpaperwork on/off — the chip above the composer controls whether the assistant answers in that conversation. Turn it off to have a team-only chat without AI; turn it back on when you want its input
What dpaperwork does in a shared conversation runs with your Role and permissions — and replies are attributed to you. It never acts with someone else's authority.
Private vs shared
- A conversation with just you stays private — delete it whenever you like
- A shared conversation is archived rather than deleted (archiving hides it from your list; the conversation stays for everyone else)
Conversations
Every Chat session lives in a Conversation — a named, persistent exchange that survives page reloads. You can:
- Start a new Conversation at any time with the New conversation button
- Return to any past Conversation from the Chats list in the sidebar — auto-grouped into Today / Yesterday / Last 7 days / Older
- Rename a Conversation by clicking its name
- Delete a private Conversation you no longer need
Conversations are also created from the ⌘K search bar — type a question there and choose "Ask AI" to open it as a fresh chat.
Voice input
On desktop you can dictate instead of typing — click the mic in the composer (or press ⌘⇧K), speak, and dpaperwork transcribes and sends it. Fillers are stripped automatically.
Tips
- Be specific about which client or record you mean when querying — "the Acme deal" beats "my deal"
- You can paste in a conversation you had with a client and ask dpaperwork to log it as an Interaction
- The onboarding exchange is saved as your first Conversation — you can always return to it and add more context
- When dpaperwork asks "Want me to save that as a decision?" — say yes. It makes future responses significantly better
- On a long reply, open the reasoning trace to see how dpaperwork reached its answer
Try it in a walkthrough
Walkthroughs follow one job end to end, in real Chat exchanges: