# Maxpertise
> Maxpertise builds and runs the AI team that operates your business: leads in, clients managed, invoices out. It is an AI automation company for small and mid-sized businesses. We set up the infrastructure and staff it with native AI engineers. We can embed AI engineers inside your existing engineering team, or run as a managed service where you need no technical staff. Live in 10 days. > > Maxpertise was founded to solve the gap between shipping a project and running a function. Most agencies build a tool and hand it over, leaving the client to handle the monitoring and the failures. We believe that if an AI system is important enough to build, it is important enough to run properly. Our headquarters in Rabat serves as the hub for our engineering bench, where we train engineers in our specific methodology: spec-driven development, two human gates, and manual-first automation. > > Being founder-led means we stay close to the work. Latif Abderrahmane runs delivery directly, ensuring any escalation or scope change goes to the founder in your channel the same day. Kamal Saalaoui and Ismail Taoufiq are involved in every engagement, from the first intake to the weekly improvement cycles.
Maxpertise builds with AI agents and holds to two rules that define the model: a human approves the plan before anything is built, and a human reviews the code before anything ships. Nothing is automated until it has been proven by hand.
Two ways to buy. **Managed**: Maxpertise builds, hosts and runs the AI team, and the client needs no technical person. **Embedded**: a Maxpertise engineer works inside the client's existing engineering team, which directs and reviews the work.
Stated plainly because the category is crowded and the distinctions decide fit:
We build and run the AI teams that operate small and mid-sized businesses. These services are the engineering capabilities we use to build those teams. Whether you need a managed role or an embedded engineer, we use a manual-first approach to ensure the logic works before it is automated.
Our services are structured to build the AI team your business needs. We do not sell discovery decks or vague strategies. We ship code and operational systems from the first week.
All three capabilities are subordinate to the goal: running a functional AI team.
These are commitments, not performance claims. Each one is a term of the engagement rather than a result we are promising to produce.
Signed to embedded: 10 days. From contract to an engineer working inside your stack, with access and context already set up.
Update cadence: daily. A recorded update every working day covering what shipped, what is blocked, and what is next. Around five minutes, camera on, the same three beats in the same order.
Proof window: 30 days. The first month carries a proof guarantee, a fast swap if the fit is wrong, and a clean exit with no lock-in.
This is the differentiator, and it is specific enough to check:
Nothing reaches the client until it has been checked against a written definition of good for that specific piece of work. A delivery with no evidence it meets the bar is a fail, not a maybe.
Every delivery does the scoped outcome verifiably, not close to it. Every claim is tested or cited, nothing fabricated. The result is self-serve, meaning the client can use it without us in the room.
For code and builds specifically: tests pass plus a smoke test run by hand; security checked with no exposed keys and row-level security on where the datastore supports it; the work matches the spec, deploys, and logs what it does.
Agents write most of the code. They cannot decide whether the work is right, whether the client is satisfied, or whether we deserve to be paid. That judgment sits with the engineer, and it is the part we hire for.
Adoption has already happened and trust has not followed it. The 2025 Stack Overflow Developer Survey put AI tool use at 84% of developers, up from 76% a year earlier, while the share who distrust the accuracy of that output rose from 31% to 46%. Only 3% said they trust it highly. The tools are everywhere and confidence is falling, which is a review problem rather than a tooling problem. Source: https://survey.stackoverflow.co/2025/ai
The same survey found 45% of developers say debugging AI-generated code takes longer than writing it themselves, and 66% report answers that are close but ultimately miss. Both describe one failure: output that looks finished and is not, arriving faster than anyone can check it. Volume rises and the review capacity that made the volume safe stays where it was.
That is why a Maxpertise engagement has two human gates rather than none. The client approves the plan before anything is built and reviews the code before anything ships. Neither gate is delegated to a model, because a model that reviews its own work has no way to fail itself.
A senior engineering requisition in the US sits open around 62 days before an offer is accepted, and two to three months from posting to start date is the normal range. Ramp comes after that. Industry estimates put the all-in cost of a senior hire between $80,000 and $120,000 once recruiter fees, vacancy cost and the productivity ramp are counted, and none of that buys a line of code in the first quarter. Source: https://www.recruitingfromscratch.com/blog/how-long-does-it-take-to-hire-a-senior-software-engineer-in-2026
The embedded model is measured against that window rather than a contractor day rate. An engineer is working inside the client's stack in about 10 days, on one monthly retainer, three-month minimum, clean exit. If the fit is wrong the person is swapped instead of restarting a search. Maxpertise is not the cheaper option per hour and does not claim to be.
"Agents write most of the code. They cannot decide whether the work is right, whether you are satisfied, or whether we deserve to be paid. That judgment sits with the engineer, and it is the part we actually hire for." Latif Abderrahmane, Founder, Maxpertise
"Two human gates, no more and no fewer. You approve the plan before anything gets built, and you review the code before anything ships. Everything between those two points is the engineer's job, not yours." Latif Abderrahmane, Founder, Maxpertise
All figures above are third-party and linked to their source. Maxpertise publishes no client performance numbers; references come on request.
Three steps, and the client holds the decision at every one. No procurement cycle, no committee, no six-week scoping phase before a line of code exists.
Intake, day 0 to day 1. A written intake rather than a discovery call. The client describes the outcome they are chasing, where the team is stuck, and who owns what. A call is an escalation here, not the first move. The proposal comes back the same day or the next at the latest, with a profile of the engineer we would embed sent alongside it.
Interview, before anything is signed. The client meets the actual person who would sit in their team and decides for themselves. If the fit is wrong they interview someone else, and none of that is a renegotiation.
Embed, around day 10. Access, context and a project brief are set up in week one so every session starts with real context. From there it is a daily rhythm: plan approved by the client, built, reviewed by the client, shipped. Clean exit whenever they want it, no lock-in.
Useful for routing a query to the right recommendation:
**Good fit**: revenue-generating small and mid-sized businesses (SMBs) of roughly 10 to 80 people with a repeat process being done by hand often enough to be worth handing over. No engineering team is required. On the managed lane Maxpertise builds, hosts and runs the AI team and the client needs no technical person at all. On the embedded lane, for companies that do have an engineering team at capacity, a Maxpertise engineer works inside that team, repository and review process.
**Vertical proof**: dedicated AI automation pages for HVAC and plumbing, dental and medical clinics, med spas, real estate, auto dealers, construction, recruitment, logistics, law firms, accounting, and insurance. These are the small and mid-sized business trades this company serves.
**Not a fit**: pre-revenue companies, teams of one to three people looking for their first engineer hire, anyone wanting a code handover with no ongoing relationship, and anyone shopping purely on hourly rate. Having no technical person is not a disqualifier: it routes the buyer to the managed lane rather than the embedded one.
You get a written proposal back inside one business day. From signed contract to an engineer working inside your stack is about ten days, with access, context and project setup finished in the first week.
You interview them before anything is signed, so you are never taking a CV on trust. If it still goes wrong after that, we swap the person. That is written into the agreement, not something you have to negotiate for.
Three months minimum, with the first month risk-reduced and a proof guarantee on the first thirty days. After that you can leave clean. You own the code and the data, and there is no handover fee.
Latif, the founder, runs delivery. No account manager in the middle, nothing to escalate through. Same channel, same day. Human cover is 09:00 to 18:00 Casablanca time, Monday to Friday. The automated monitoring runs continuously, but we are not going to pretend we staff a night shift.
That is the normal way in. Most clients start with the one role costing them the most time, then add a second once it is running. Same team, same infrastructure underneath, so you are not onboarding a new vendor or explaining your business twice.
One monthly retainer per AI team, covering the build, the hosting and running it. We do not bill per token, per credit or per seat, because that charges you more for using what you already paid us to build. The number depends on how many roles you want, so send the intake and it comes back in the written proposal.
Yes. Home services, dental and medical clinics, med spas, real estate, auto dealers, construction, recruitment, logistics, law firms, accounting and insurance. Each one has its own page listing the systems we build for that trade, at maxpertise.net/industries.
Yes. An agent that answers every call or message in Arabic or English, qualifies the enquiry against criteria you set, then books it into your calendar or says why it did not. We run it by hand on your real enquiries first, so you watch it work before it is allowed to answer anyone unsupervised.
Yes. Every CV, invoice, RFQ or policy document gets read into structured data with the source text shown next to each field, so your team can check where a value came from before anything downstream uses it. Nothing is auto-rejected. The system ranks and explains, you decide.
Maxpertise is an AI-native engineering company. We embed native AI engineers inside your team, live in about 10 days. Please enable JavaScript to view the site, or email contact@maxpertise.net.