A mid-sized company should get an AI consultant who works at its scale: someone who finds the few workflows that eat the most hours, ranks them with numbers, agrees with the owner or COO which comes first, and leaves a plan your team can run. You should not get a programme built for a company with its own AI team.
#Key takeaways
- At 200 to 2,000 staff, most companies have no AI team, a handful of heavy workflows and decisions that sit with an owner or a small leadership group.
- Proposals written for enterprises assume a data team and a programme office; ask for one scoped first workflow instead.
- The useful output is a ranked list of workflows with numbers and an owner, not a maturity model.
- The model is rarely the hard part; data, process and who approves the output usually are.
- Ask who builds the first workflow and who runs it afterwards before you sign anything.
#What does AI consulting look like at your size?
Start from your week. Finance closes the month in SAP or Odoo with a stack of Excel reconciliations. Customer service answers the same questions across email and WhatsApp. Sales updates HubSpot or Salesforce when someone finds the time. It runs Microsoft 365 with a small, stretched team. There is no data scientist.
In that setting, most of the consulting work is a data and process problem. The model is the easy part.
#Why do enterprise-sized proposals miss for mid-sized companies?
Large-firm proposals often assume a dedicated internal AI team, a data budget and a programme office to absorb the recommendations. A mid-sized company has none of those, so the plan stalls after the readout. The scope is too wide and nobody inside owns it.
Watch for a proposal that jumps to a multi-agent system or a custom platform before anyone has looked at your processes. Gartner predicts over 40% of agentic AI projects will be canceled by the end of 2027, and warns about agent washing, where vendors relabel chatbots and rpa as agents (Gartner, june 2025).
Simpler is usually right. Anthropic's own guidance is to start with the simplest approach and add complexity only when it is needed (Anthropic, december 2024).
#What should a consultant deliver to a company your size?
- Interviews with the people who run the heavy workflows, not only with their managers.
- A ranked list of a few workflows, each with volume, the cost of the current process and the cost of a mistake.
- One recommended first build, scoped small enough to run with real users, with its hosting region.
- A named owner on your side for each workflow, and a note on who approves AI output.
- A plan your team can follow without the consultant in the room.
For how a full plan should read, see what a good AI roadmap for a mid-sized company contains.
#What should you ask before hiring an AI consultant?
- Have you put a workflow into production with real users at a company of our size? Show us.
- Who exactly does the work, and are they the people in this meeting?
- What do we receive at the end, and can we use it with another builder?
- Where will our data be processed, and who decides that?
- Who runs the system after it is built, and who answers when it fails?
Ask for production experience, not demos. A longer list is in how to choose an AI consulting company.
#When do you not need an AI consultant?
When the first step is obvious and an existing tool covers it. If the real problem is that staff ignore the Copilot or ChatGPT licences you already have, training is the answer, not consulting. If you have a capable it lead and one clear workflow, a builder may be enough. do you need an AI consultant? walks through the cases.
#Where does Stennir fit for a company your size?
Stennir works with operations-heavy companies of 200 to 2,000 staff, first in the UAE and then Saudi Arabia. AI consulting starts with the AI Opportunity Audit: five ranked workflows, each with a cost and a payback case, and one recommended build with its scope and hosting region. An executive workshop agrees an owner per workflow.
The report is yours whether or not you build with Stennir. If you go ahead, the same people build the first workflow in AI implementation, in Arabic and English, with a named person on your side approving AI output.
We are
Stennir scopes AI work to the team you have: a few ranked workflows, one first build, owners on your side and a plan you can run.
We aren't
Stennir does not sell an enterprise programme that assumes an in-house AI team, a data department and a programme office you do not have.
#What should an owner or COO do next?
List the three workflows that cost your teams the most hours and the systems they run on. Then take the free readiness assessment for a quick read on where you stand. See AI for operations for typical starting points, or book a 30-minute call to test your list.
#Common questions
Is my company too small for an AI consultant?
Not if a few workflows take a large share of your team's hours and the people involved can explain them. Repetition and volume matter more than headcount. A company of a few hundred staff with heavy invoice, order or ticket flows often has clearer first candidates than a larger one with scattered work.
What do you need from us to start?
An owner or COO who can decide, access to the people who run the heavy workflows, exports or read access to the systems involved, and a finance contact who can confirm what the current process costs. You do not need an AI team or a data department to start.
What stays with us when the engagement ends?
The audit report, the ranking, the cost cases and, if you build, the accounts, the data and the off switch. Stennir works inside accounts you own, so stopping does not mean losing access to your own systems or your own record of approvals and logs.
Does it work in Arabic?
Yes. Stennir works in Arabic and English, including Gulf dialect messages and Arabic documents such as invoices and contracts, and other languages on request. Language is tested on your own documents during the audit, because it often decides which workflow is ready to go first.
