Answers from your own data. With the source attached.
We build assistants that answer staff questions from your sales, policies and documents, inside the chat apps your team already uses. Each answer shows where it came from, and when your data doesn't cover a question, the assistant is built to say so and pass it to a person.
This is for you if
Three signs it is time.
- Managers message the office for yesterday's sales, then wait for someone to look it up.
- The same questions about policies, approvals and how-tos land on the same few people every week.
- A job like captioning video or sorting files takes hours by hand and follows the same rules every time.
What changes
Answers where people already talk
Staff ask in the messaging app they already use and get yesterday's sales, a policy or the right approver as a reply.
Answers you can check
Each reply links to the report, document or record behind it, so a manager can check a figure before acting on it.
Busywork that runs on its own
Repetitive jobs, like captioning a video word for word from its script, run by themselves. A person reviews the result instead of doing the work.
What's included
The usual scope. Every quote says exactly what is in and what is out.
What it reads
- Live sales from your point-of-sale systems
- Policies, handbooks and how-to documents
- A knowledge base your team can correct and add to
- Spreadsheets and shared drives
Where it lives
- Inside your team's messaging apps
- A phone app for alerts and reports
- A web portal to teach, approve and broadcast
- Desktop tools for video and file jobs
Controls
- Answers limited to what each person is allowed to see
- The source shown with every answer
- A log of every question and reply
- A hand-off to a person when the answer isn't in your data
- Approval steps before anything is sent or changed
How it runs
The general process, shaped for this kind of job.
- 01
Collect the real questions
We gather the questions people actually ask, from chat threads, email and the office, and sort them by how often they come up and how much a wrong answer would cost.
- 02
Connect the sources
We connect the systems and documents that hold the answers, carrying each person's access through, so the assistant never tells a store manager something they couldn't already see.
- 03
Test against known answers
Before staff see it, we run it through a set of questions with checked answers. We keep that set and re-run it after every change.
- 04
Start with one team
One team or region goes first. We read the logs, fix wrong or missing answers in the knowledge base, then open it up.
More of this kind of work. From the index.
Questions
Asked on first calls about this.
What stops it making things up?
It answers from your sources, shows them with each reply, and is tested against questions with known answers before and after every change. No assistant is perfect, so we design for checking: the source is one tap away, and anything it can't find goes to a person.
Where does our data go?
Your sales and documents stay where they are. The assistant fetches only what it needs for each question and sends that to an AI model to write the reply. Before anything is connected, we show you which provider that is, what their terms say they keep, and the alternatives.
Can it do things, or only answer questions?
Both, within limits you set. The team assistant handles approvals, campaigns and broadcasts as well as questions. Anything that spends money, changes records or messages customers waits for a person to approve it.
Do we need lots of clean data first?
No. Most assistants start from what you already have: a point-of-sale system, a handbook, a shared drive. The bigger job is usually tidying documents so there's one current version of each answer.
Goes well with
Services
Industries
Want this for your business?
Tell us what you have today and what is getting in the way. A person reads it and replies within two business days.