# An AI consulting company that shows where AI pays off in your business

AI consulting is the work of deciding which processes in a company are worth automating with AI, in what order, and whether to build or buy. Goldenroz runs a fixed-price assessment for companies that have budget for AI and no agreed first project. You get a ranked list of use cases, a build-vs-buy recommendation for each, and a pilot plan with a price. The assessment is quoted after a 30-minute scoping call, and builds that follow start at $25,000.

**Fixed-price assessment, quoted after a scoping call**

Updated: 2026-11-13. By Shai Goldenberg.

## Who this is for

- CEOs and COOs who have been asked for an AI plan and want one grounded in their own processes and numbers.
- Operations leaders with several candidate processes and budget for one or two.
- CTOs who need an outside engineering view before committing a team to an AI build.
- Companies comparing vendor tools with a custom build and unsure which costs less over two years.

## What is included

- **Process interviews:** Sessions with the people who do the work today, to record steps, volumes, exceptions and where time goes.
- **Data and system review:** A check of where the data lives, who may access it and how an AI system would reach it: API, database or export.
- **Use-case ranking:** Each candidate scored on value, feasibility and risk, with the reasoning written out so your team can challenge it.
- **Build-vs-buy recommendation:** For each top use case: an off-the-shelf tool, a custom build, or a wait, with the cost drivers of each option.
- **Risk and security notes:** What data the model would see, which actions need human approval and what has to stay on your own servers.
- **Pilot plan:** A scope, a success measure, a timeline and a fixed price for the first build, written so any competent team could deliver it.
- **Readout:** A working session with your leadership to walk through the ranking and settle the first project.

## How the work runs

1. **Scoping call (30 minutes):** We agree which departments and processes are in scope. You get a fixed price for the assessment.
2. **Discovery (First stage):** Interviews, process walkthroughs and a review of the systems and data behind each candidate.
3. **Ranking and build-vs-buy (Second stage):** Candidates are scored and the top ones get an option comparison with cost drivers.
4. **Readout and pilot plan (Final stage):** We present the ranking, answer objections and hand over a pilot plan with a fixed price.
5. **Pilot (Optional):** If you build with us, a prototype of a focused agent typically takes one to two weeks. See [AI agent development](https://goldenroz.com/ai-development/).

## How an AI consulting firm should rank use cases

Most AI consulting services produce a long list of ideas. A list is not a decision. We score each candidate on four things: how often the task happens, how clearly a person can explain the rules, whether the data is reachable, and what a wrong answer costs.

High volume with clear rules and cheap errors goes first. Drafting replies to routine support tickets fits that pattern. Low volume with expensive errors goes last or gets a human approval step. Approving refunds above a threshold fits that one.

The ranking is a table your finance lead can read. Each row shows the hours or cost at stake today, the assumptions behind the estimate, and what would have to be true for the project to pay back.

## Build, buy or wait

Buying wins when the process is common across companies and a vendor already sells it. Meeting notes, generic support deflection and code assistance are usually bought. Building custom AI solutions wins when the process depends on your own data, your own rules or systems a vendor cannot reach.

The comparison covers more than the license fee. We look at per-seat or per-resolution pricing at your volume, integration work the vendor leaves to you, who holds the data, and how hard it is to leave. A tool that is cheap at 1,000 conversations a month can cost more than a custom build at 50,000.

Waiting is a valid result. If the data is not reachable or nobody owns the process, the recommendation says so and lists what to fix first.

## AI workflow automation: consultant or automation agency

An AI automation agency typically connects existing tools with no-code workflows. That works for simple handoffs between apps. It gets fragile when a workflow needs permissions, retries, an audit trail or access to an ERP.

As an AI automation consultant, Shai looks at each workflow as an engineer. Some should stay as plain code with no model at all. Some need one model call inside a fixed workflow. A few need a real agent that plans and takes actions. The assessment says which is which, because the cost differs by an order of magnitude.

## What backs this up

- **Advice from the person who builds:** [Shai](https://goldenroz.com/about/shai/), our CTO, runs the assessment. He has shipped a production WhatsApp ordering bot on Claude and a support assistant with a knowledge base.
- **Enterprise data experience:** He owns an on-prem MCP gateway with 13 connector types that lets Claude or ChatGPT answer from ERP, CRM, SQL and Microsoft 365 systems read-only.
- **A plan you can take elsewhere:** The ranking, the comparison and the pilot plan are yours. Nothing in them requires you to build with us.
- **Published build pricing:** Builds start at $25,000 and are fixed price per phase, as shown on the [pricing page](https://goldenroz.com/pricing/).

## What it costs

The assessment is a fixed-price engagement. The price depends on how many departments and processes are in scope and is set after a 30-minute scoping call.

- AI assessment: Fixed price (Quoted after a 30-minute scoping call)
- Pilot or first build: From $25,000 (Fixed price per phase)
- Larger AI platforms: $100,000 and up (Scoped per project)
- First consultation: $0 (Free 1-hour call)

## Who it is not for

- Teams that already know what to build. Go straight to [AI agent development](https://goldenroz.com/ai-development/).
- Companies looking for a slide deck on AI trends. The output is a decision about your processes.
- Organization-wide change management or staff training programs. For engineering teams, see [Claude Code training](https://goldenroz.com/claude-code-training/).

## Questions buyers ask

### What does an AI consulting company do?

It helps a business decide where AI is worth the money and how to get it into production. At Goldenroz that means interviewing the people who do the work, reviewing data and systems, ranking use cases, recommending build or buy for each, and writing a pilot plan with a fixed price.

### How much do AI consulting services cost?

Our assessment is a fixed price quoted after a 30-minute scoping call, and it depends on how many processes are in scope. Builds that follow start at $25,000, fixed price per phase. The first consultation is a free 1-hour call.

### What do we receive at the end?

A ranked table of use cases with the estimate behind each, a build-vs-buy recommendation for the top ones, security and risk notes, and a pilot plan with scope, success measure, timeline and price. All of it is written to be used by any delivery team.

### Do we have to build with Goldenroz afterward?

No. The deliverables are yours and are written so another team can execute them. If the right answer for a use case is an off-the-shelf tool, the report says that.

### How is this different from an AI automation agency?

An automation agency usually starts from a toolset and looks for workflows to connect. We start from your processes and decide whether each one needs a purchased tool, plain code, a single model call or a full agent. The person advising is the engineer who would build it.

### Is our data safe during the assessment?

We review how systems are structured and how data would be accessed. We do not need to copy production data out of your environment. Access is read-only, limited to what the scope requires, and agreed in writing before work starts.

### What makes a good first AI project?

A task that happens often, follows rules a person can explain, uses data that is already reachable, and has a low cost when the output is wrong. Projects that fail usually start with unclear ownership or data nobody can reach.

---
Page: https://goldenroz.com/ai-consulting/
Book a free one-hour call: https://calendly.com/daniarozin/daniel-rozin-consultation-1-hour-clone
