Your company's AI department.Without hiring one.
You are an organization or a professional firm that wants its people to actually work with AI. We come in as a department: train your people, connect the AI to the systems you already run, and stay to maintain it. Subscriptions in your name. The know-how stays with you.
Without hiring, without buying a platform, and without your data leaving the building.
- Organizations and professional firms
- Your subscriptions, your keys
- Data never leaves the company
Which orders from this week aren't in Priority yet?
3 WhatsApp orders pending approval:
- Shufersal · 240 units · 12 Sep
- Rami Levy · 90 units · missing price
- Yochananof · 120 units · 14 Sep
I saw from the inside how corporations roll out AI, and what breaks along the way.
I founded Solviz to give every organization what corporations have in-house: an AI department.
Omer Lapidot
Ex-Wonderful, where he saw up close what happens to enterprise AI projects after the demo. Before that, an operations commander in the Israeli Air Force. Believes you never sell an organization an agent before its people can work with the tool.
Three ways to start. They all end in the same place: your people working with AI every day.
We don't sell a platform or an agent. We join the organization as a department, with a clear hours quota and a price quoted upfront.
Mapping & prioritization workshop
One day with your department heads. You leave with ten use cases, and a number next to each.
- Prep interviews with 3 to 5 managers
- A full day on site, laptops open
- A ranked list: what to do first, and why
External AI department
Monthly retainer in hours. Every AI question in the company comes to us, not to your IT person.
- Validation of what was already built, before building more
- A new department scoped every month
- Building together with the team, until it builds alone
One system that works
One process, one engine. Delivered in a month, maintained for six.
- Scoping session and a signed scope document
- Delivery, training and rollout with the team
- At the end, the code and data transfer to you
Didn't leave the workshop with five use cases and a number next to each? You don't pay for it.
And the external AI department has no minimum term. A month that brought no value, we don't continue. That is how we make sure every month has to earn the next one.
Three steps, and one that never ends.
- 01
Pick the system
Claude, ChatGPT or Copilot, based on the systems you already run. The subscription in your name, from the vendor.
- 02
Connect the systems
ERP, CRM, email, calendar and WhatsApp. Through MCP when an official connector exists, built by us when it doesn't. All inside your environment.
- 03
Train the people
Groups of five, each department on its own examples. Not a lecture in a hall.
- 04
Stay
Answer questions from the floor, build the next one, recalibrate. Every month, with a report.
of enterprise AI pilots produced no measurable result
MIT NANDA, The GenAI Divide, 08/2025
more projects reach production when someone outside is accountable for it working (67% vs. 33%)
MIT NANDA, The GenAI Divide, 08/2025
“Most organizations have already paid for AI. Few got anything back. That gap is what we close.”
Traditional organizations. With an ERP. Without a dev team.
That is our definition of a client. Have an R&D department? Probably not a fit. Have a core system everyone depends on and one IT person? That is why we started the company.
Manufacturing
Shift handovers and faults nobody documents.
An automatic shift brief, and a bot answering procedure questions from the documents.
Food & beverage
Orders arrive by email and WhatsApp and get keyed in by hand.
Order intake straight into the ERP, with a human approving every line.
Distribution & logistics
Complaints discovered only after the customer is already angry.
Real-time exception tracking, and a drafted customer reply waiting for approval.
Construction & surveying
Public opportunities surfacing late, or not at all.
An engine that pulls the data weekly, scores it, and writes the outreach.
Automotive & telematics
Bots that were built and stalled, and a backlog waiting on one IT person.
First, validation of what exists. Then an AI department building with every team.
Retail
Thousands of frontline questions, the same answer over and over.
Answers for field staff on WhatsApp, from the procedures, no new app.
Professional services
Knowledge lives in people, not in a system.
Meeting prep, call summaries into the CRM, and document drafts in the firm's format.
Public sector
Inquiries and documents that take weeks to get through.
Triage, drafted replies for approval, and summaries of long documents with sources.
The same architecture for every client. Adapted to the systems you already run.
An independent agency. Not affiliated with any vendor, and we connect to all of them.
Clients who went through it with us.
Within a month we had an engine handing sales a weekly list with the message already written. Instead of analysis in someone's head, there is a work list.
We came with stalled bots. We got a validation report, priorities, and a team that sits with us until we build alone.
The workshop pulled one use case out of every department, with a number next to it. That is what moved it from management to the floor.
What we learned this week in the field.
What we get asked on the first call.
A workshop at a fixed day rate. A retainer in a monthly hours quota, no minimum term. A project with a one-time fee plus monthly maintenance. It usually comes to less than one in-house AI hire's salary, with no recruiting and no dependence on one person. You get the numbers in a written two-page proposal, before we start.
Got a process that's blocking the organization? Let's open it.
30 minutes. No selling, no slides. We map the first step together and check whether there is ROI here.
- The procurement manager doesn't type orders. She approves them.
- The shift manager gets a brief at 6:50, before he walks onto the floor.
- The IT person answers questions. He is no longer the bottleneck.