Motive Chat
The workspace your team opens every day. Chat, board, CRM, dashboard, hub, settings. Your assistants are members of it, and every capability they hold is set to on, ask first, or off. Client seats onboard this fall.
See the workspaceThree things you can buy, separately or together.
Motive Chat is the workspace your team opens. The shared brain is what your business knows, in plain files you own. The custom software is what we build per role, on your machines.
Two custom operator seats, installed inside two weeks of the go-ahead.
Three products. One, two, or all three. Each one stands on its own.
The workspace your team opens every day. Chat, board, CRM, dashboard, hub, settings. Your assistants are members of it, and every capability they hold is set to on, ask first, or off. Client seats onboard this fall.
See the workspaceOne private knowledge layer for your business. Plain files and open standards, so a correction you make today is still in force next quarter and none of it is trapped inside one company's product.
See what it holdsOperating systems and agents built one per role, on your machines and your subscription. Websites and web platforms when you need those built too.
See what we buildThe systems we build post into Motive Chat. The shared brain is what they read and write. Chat is where you see all of it, talk to it, and approve what it does.
Not a chat window. Six surfaces your team works in, with your assistants as members of the room.
Chat, board, CRM, dashboard, hub, settings. An assistant here is a member with its own profile in the room. You @mention one and it goes to work. What it proposes comes back as a card inside the thread you were already reading, and every assistant post is checked before it appears. If the check fails, the post does not happen.
A workspace does not get every surface. The dashboard, the hub, the CRM and the board are each a switch we set per client, and the module list is growing.
Today
Morning brief.
Build: green on main, all checks passing. The overnight job retried once and recovered on its own.
Clients: one invoice is queued for approval and one proposal is waiting on review. The onboarding doc is drafted.
Calendar: clear until the 2pm sync. I blocked 90 minutes this afternoon for deep work.
solid. queue the invoice and push the doc to review
Done. Invoice queued, doc is in review. I will flag both when they clear.
@assistant rerun the brief with the calendar fix
Message #briefing
The assistant's mark is set in a different face from a person's, and the sender cannot change that. You always know which one you are reading.
Every capability an assistant holds is on, ask first, or off. Ask first is not an instruction we ask a model to follow. The function is simply not handed to the assistant when it starts work, so there is nothing for a clever prompt to talk it into and refusal never depends on the model behaving well under pressure.
That is the honest answer to why not just use a chat window with plugins, where every tool it has been given is callable on every turn and judgment is the only thing in the way.
Some of these we hold. Sending from your account is held at ask first or lower. Deleting mail, changing your forwarding rules, and moving money are held off with no path to on. Nobody can raise those from inside the product, including us. Changing one is a code change, not a setting.
Every capability your assistants hold is listed here. A capability set to ask first is left out of the toolset when an assistant starts work, so it cannot do the thing at all. It can only post a card here and wait for you.
6 on · 2 ask first · 3 off
google-workspace
read your calendar
CALENDAR.READ
add and move events on your calendar
CALENDAR.WRITE
read the files you share with us
DRIVE.READ
read your inbox
MAIL.READ
write email drafts and leave them in your drafts folder
MAIL.DRAFT
send email from your account
MAIL.SEND
Motive holds this at ask first or lower: sending is always yours to approve
delete email and empty the trash
MAIL.DELETE
Motive holds this off: a delete cannot be taken back from here
change your mail settings, including forwarding and filters
MAIL.SETTINGS
Motive holds this off: a forwarding rule is silent, permanent, and invisible from the inbox
quickbooks
read your invoices and their status
INVOICE.READ
Every connector below is a declared list of what your assistants may do with it, written in plain English. Open one to grant it, connect it, and set what it may do.
Productivity
Google WorkspacePLANNED
Gmail, Calendar, Drive, Docs and Sheets for a business running on Google.
9 on · 4 ask me first · 4 off
Microsoft 365PLANNED
Outlook, Calendar, OneDrive, Word, Excel and Teams for a business running on Microsoft.
12 on · 4 ask me first · 7 off
SlackPLANNED
Read the channels it is invited to, post as itself, and keep canvases up to date.
9 on · 3 ask me first · 1 off
NotionPLANNED
Read and add to the pages and databases you choose to share with it.
7 on · 2 ask me first · 1 off
Finance
QuickBooks OnlinePLANNED
Customers, invoices, bills and reports from the books you already keep.
7 on · 5 ask me first · 2 off
CRM
HubSpotPLANNED
Contacts, companies and deals, with notes and tasks kept current for you.
9 on · 3 ask me first · 3 off
Commerce
StripePLANNED
Customers, invoices, payments and balance, read from your own Stripe account.
7 on · 6 ask me first · 3 off
SquarePLANNED
Sales, customers, catalog, stock and invoices from your Square account.
11 on · 7 ask me first · 3 off
Eight connectors are written and declared. The policy layer, the manifests and the approval ladder are built. Every card reads PLANNED until we wire the handshake to your own account at install, which is exactly what it should say until we do.
The rule we build by: nobody migrates off a tool that already works. If your business runs on Square or Shopify or QuickBooks or Jobber, that stays your system of record. We build around it.
Your tokens, your mailbox access, and the runtimes that hold them stay on your own machines. They never enter the shared database, the hosting environment, or the repository. What crosses into the workspace is only what your own runtime chooses to post: a message, a brief, a rendered digest.
Motive Chat runs on our own network today, with client seats onboarding this fall. It is installed for your team, not sold as a signup.
One private knowledge layer per client. Plain files, open standards, and corrections that stick.
A raw model starts blank every morning. The brain holds your chart of accounts, your close calendar, your vendors, and the way your team actually works, so nobody re-explains the business at the start of every session. Update it once and every seat is current.
Correct a chat window and it forgets by tomorrow. Here the correction gets written down and the next release has the rule in it. Six weeks of one finance lead's corrections produced five releases, and most of the changes were authored from his own logs rather than from our assumptions. Anyone starting from scratch today is six weeks behind, and the gap widens every week.
It is plain files and open standards, not a wrapper around a single vendor's model. If the pricing, the terms, or the product changes, the thing you paid for is still yours and still works. We run our own agency on this same layer, which is how we find its edges before a client does.
An owner and a finance lead have almost nothing in common in what they need from software, so they get different systems. On the live install, the owner seat runs inbox triage, replies drafted in his own voice, a daily briefing, the calendar, and file organization. The finance seat runs daily sales journals, semimonthly flash reporting, the full payroll set, weekly cash flow, and tax prep. Both went live inside two weeks of the go-ahead, on the client's own machines and the client's own AI subscription.
It runs on a schedule without being asked, and it says when it does not know instead of smoothing it over. More than forty consecutive daily runs, every reconciliation check clean. Along the way it surfaced a calculation defect in a recurring filing and a set of payment batches that had never settled, both of which the manual process had carried for months.
The gates were designed before the features.
It drafts email and never sends. It never moves money, never files with a government, never submits payroll, never deletes. It started read-only and earns write access one lane at a time, on the client's explicit approval. Every one of those limits is enforced in the software, not in a policy document.
Full rebuilds and migrations, online reservation and fitting-request flows, SEO, security headers, and CI. On the build running now we inventoried every indexed URL on both sites first, so nothing that ranks today is lost in the move. A certificate error that had been served to every visitor who typed the www address gets fixed in the cutover at no extra charge, because we found it in the audit.
We also built a payments platform for construction draws: custom agents, contract review, and operator dashboards. Delivered, and running all summer.
You own it outright.
It runs on your machines, on your subscription, with your data in your own accounts. Nothing routes through a server we own. No migration, no per-seat platform fee, nothing to be locked into. What you leave with is infrastructure that keeps working after we step back.
Before we wrote a line of code, this client was being sold a point-of-sale migration and had no independent way to judge it.
We sat in the vendor meetings. We checked the API claims against the published documentation instead of the sales deck. One claim made in those meetings was simply wrong. We wrote up what we found, the reasoning behind it, and the condition under which our own recommendation should be reversed. The decision is still theirs, on their timeline.
We were not paid for any of it. It is the reason we were trusted with everything after.
A free evaluation became a paid build, then a website contract, and is becoming a company-wide platform. Nobody was asked to sign for all of it up front.
Everything is working great currently. I've got a few of my regular tasks pretty dialed: daily sales, bi-monthly flash report and payroll are the main ones. It's also been super helpful with various one off analyses.
One client's finance operator started building his own routines on top of the system, without us. That is the point.
Five locations, two seats built separately. The owner's runs inbox triage, drafted replies, and a daily briefing. The finance lead's runs daily sales, the flash report, payroll, cash flow, and tax prep.
Two rebuilds for the same retailer, every indexed URL inventoried, with the full redirect map. Online reservations and fitting requests, an SEO layer, security headers, and CI. Next.js on Vercel.
A payments platform for construction draws. Custom agents, contract review, and operator dashboards. Delivered and running all summer on Next.js and Supabase.
A multi-seat install for a manufacturer and distributor. Per-seat workflow builds on one shared brain, so an item update stops cascading by hand.
The workspace. Chat, board, CRM, dashboard, hub, settings. Assistants are members and every post is checked before it appears. Running on our own network today, with client seats onboarding this fall. Installed for your team, not sold as a signup.
The operator dashboard we run the agency on. It is also the starting template for every client system we build.
A seat for your team, one brain behind it, and an agent for every job you do twice.
Book the call
Motive Chat, the shared brain and the agents all run our own agency first, on our own console. What we install for a client is the same architecture, shaped to their business. The run counts here come from the finance seat we built for a five-location ski and bike retailer.
15 minutes. We map the jobs you repeat and say the same day if we can take them.
We wire your tools and stand up the seats. Two went live inside two weeks.
Your team asks. The work gets done and the record gets written.
You correct it, we bake the rule in. It reports its own health and says when a machine goes quiet.
The first call is 15 minutes. If we can't save you real time, we say so that day.
No. Your team asks in plain language. The technical layer stays ours.
George and Dexter. Not a contractor team.
Yours. You keep your own account and your own billing, on the plan we specify at install. We carry no model-spend markup, and you are never inside ours.
It stays in your accounts. We don't pool client data. You hold the keys.
A setup fee to build it, then a monthly to run it. Websites and platform builds are quoted per project. You get the number on the first call.
Cancel any time. You keep your data and the integrations. Code ownership is set per contract.
We are George and Dexter, two students at the University of Utah. We started Motive because we kept watching good operators lose their week to work that software should have handled.
We design, build, and run every system ourselves. You talk to the people who wrote your code, not an account manager or a contractor team.
If you need to change something, you call the person who built it.
Not for businesses that want a vendor. For operators who want a system they own.
Motive LLC, Salt Lake City. Every system on this page was built by the two of us.
15 minutes. No deck. We map what's eating your week and tell you on the call whether we can take it.