Ask your data. Draft the reply. Move on.
A copilot that summarises leads, drafts follow-up emails and answers plain-language questions grounded in live data from every module — not a generic chatbot.
People & Support · Copilot, lead summaries & @AI chat
A copilot that has actually read your data
A generic chatbot bolted onto a CRM can summarise text you paste into it. That is not the hard part. The hard part is answering "which of my open deals went quiet this month, and what did we last promise them" — a question that needs the pipeline, the WhatsApp threads, the quotations and the support tickets at once.
Because every Dolfin360 module writes to the same database, the AI Assistant can answer from live records rather than from a vector index of last month's export. Ask in plain language, get an answer grounded in what is actually in the system, with the records it drew from linked so you can check its work. Generation is metered at a published credit rate, so the cost of asking is never a surprise.
Grounded in live records
Answers are drawn from current module data and cite the records they used, so you can verify rather than trust.
Drafts in your context
Follow-up emails, WhatsApp replies and meeting summaries drafted from the actual deal history, not a blank prompt.
@AI anywhere there is a note
Mention @AI in a note on any record — a deal, a ticket, an employee — and get an answer in the thread where the question came up.
Ask your pipeline a question
Plain-language queries answered from live module data, with citations.
Type the question the way you would ask a colleague: which deals slipped a stage this quarter, which accounts have an open ticket and a renewal inside sixty days, what did we quote this customer last time. The assistant resolves it against current records and shows the ones it used.
That citation list matters more than the prose. It is the difference between an answer you can act on in a forecast call and one you have to go verify by hand — and it is only possible because the assistant is reading the same rows your team is editing.
- 01Plain-language queries — no query builder, no saved-report hunting
- 02Linked citations — every answer shows the records behind it
- 03Cross-module scope — deals, tickets, threads and employees at once
What is inside the AI Assistant
Plain-language queries resolved against live module data.
Every answer names the records it was drawn from.
Generated from real activity, not manual notes.
Follow-ups written in the context of the deal.
Ask inside a note on any record and get an in-thread answer.
Generation billed at a transparent credit rate per call.
Summaries and drafts where the work happens
Lead summaries, reply drafts and @AI mentions inside any record.
Open a lead that has been running for three months across two reps and a support escalation, and the summary is already there: what was promised, what objections came up, where it stalled. It is generated from the record's own history rather than from someone's notes.
Drafting works the same way. Ask for a follow-up email and it arrives with the right names, the right last-quoted number and the right open question — a starting point you edit, not a template you fill in.
- 01Lead & account summaries — generated from real activity history
- 02Reply drafting — email and WhatsApp, in deal context
- 03@AI in any note — answers in the thread where you asked
How a question becomes an answer
Ask
In the assistant, or by mentioning @AI in a note on any record.
Resolve
The question is answered against live rows across the modules you licence.
Cite
The records used are linked, so the answer is checkable rather than trusted.
Act
Turn the answer into a draft email, a task or a stage change in place.
Metering, in the open
Platform facts and published rates — not customer-measured outcomes.
“The citations are the feature. An answer we can check in two clicks gets used in a forecast call; one we cannot check gets ignored.”
Ask your own data a question
Turn on the AI Assistant and try it against your live pipeline for 14 days.
AI Assistant questions
From live rows in the modules you licence — deals, tickets, threads, employees — not from a periodic export or a separate index. Every answer links the records it drew on so you can verify it rather than trust it.
Each generation consumes 5 credits at the published rate. Your plan includes a monthly credit grant, and usage is visible per user and per module so there are no end-of-month surprises.
Live: the copilot, lead and account summaries, email and message drafting, and @AI mentions inside any record note. On the roadmap and clearly marked as such: predictive deal scoring, autonomous multi-step workflow agents and voice-note transcription.
Try AI Assistant free for 14 days
No credit card. It shares data with every other module the moment you turn it on.