AI strategy and delivery
A fractional AI director.
At CloudTech, your AI plan is written by the people who will build and run it.
A CloudTech AI director owns the subject on your side: measures where the company loses time and money, picks the changes that pay off and holds each one to a measurement. Our engineers build the integrations, automations and assistants — with no handover to yet another contractor.
The review starts with five questions and takes about two minutes.
AI is already in your company. Nobody is accountable for it.
- People use AI chat tools on their own, with company data and no agreed rules.
- Several teams are trying different tools, but nobody has checked which ones actually shortened the work.
- Your CRM and ERP vendors are adding AI features to the subscription, and you cannot tell whether they are worth paying for.
- The board expects an AI plan, and nobody on the team can give it even a few days a month.
All four have the same cause: AI has no owner. A fractional AI director becomes that owner.
From measurement to model.
The drawing at the top of this page is our method: four layers, each resting on the one below. AI sits on top, because without a baseline, a working process and data it would have nothing to rest on.
Measurement
We establish the baseline: how long an enquiry takes, what a mistake costs, where work waits for a person. Without it, no change can be proven later.
Workflow
We map how a task moves through people and systems. Often the first improvement is removing a step, not adding a model.
Data and integrations
We check where data lives, who owns it and how systems connect. No model can replace this foundation.
AI layer
Only then do we choose a model, assistant or automation, and connect it to the systems that already work.
After each layer you get a short summary and decide whether we continue.
Every change has an acceptance criterion.
Before we build anything, we write the change down the way an engineer writes a requirement: what we measure today, what changes, how we will know it worked and who is responsible. After launch, we come back to the specification and compare.
- Process
- Replies to quote requests from the website form and email
- Baseline
- An average of 26 hours from enquiry to first reply
- Change
- An assistant classifies the enquiry, pulls the customer history from the CRM and drafts a reply for a salesperson to approve
- Acceptance criterion
- A first reply within 4 working hours for 8 in 10 enquiries
- Owner
- Head of sales
- Result check
- 30 days after launch
An illustrative example. We set the criteria for your company from the baseline measured at the start.
An AI director with engineers behind them.
Plan and delivery in one team
The same team assesses an idea, prices it and builds it. Nothing gets lost in a handover between an adviser and a contractor.
Running costs counted from the start
We run systems on AWS and Google Cloud every day, so we count not only the launch of a change but also what it costs after a year in production.
Tools that fit your systems
We start from what you already have: licences, CRM, ERP and cloud. We propose a new tool only when the current ones are not enough.
Data under control
For every change we agree which data may reach an external model and which stays in your infrastructure. We take GDPR and the AI Act into account.
Work in a monthly cycle.
A steady rhythm makes progress something you can check, not just read about. Every cycle ends with acceptance and a report.
Selection
We check the results of earlier changes and use the measurements to pick the next one.
Build
Our engineers build the change and test it on your data before it reaches your team.
Acceptance
The change goes into everyday use once it meets the criterion in its specification. We show your team how to use it.
Report
At the end of the month you get a summary: what changed, what it achieved and what we propose next.
Getting started
A baseline of selected processes, a list of possible changes with estimated gain and cost, and a first specification ready to build.
Ways to work with us.
Start with the free review. You commission each next stage separately, once you know what it will deliver.
AI review
Free
- Five questions online and a call with someone who runs deliveries
- A first list of processes where AI could help
- A clear answer on whether and where to start
Delivery map
One-off project · priced after the review
- A baseline measurement in selected teams
- Change specifications with acceptance criteria
- A delivery order with build and running costs
- A document you can deliver with us or another contractor
Fractional AI director
Billed monthly
- A dedicated person accountable for your AI programme
- A monthly cycle: selection, build, acceptance, report
- Oversight of AI vendors and tool costs
- CloudTech engineers to deliver the changes
Typical deliveries
Each one is priced upfront, from its change specification.
- Assistant for company knowledge and documents
- Handling enquiries from forms and email
- AI in your CRM or ERP
- Reports generated from company data
- Reading documents and invoices
- Internal tools with a chat interface
Free AI review.
Answer five questions. We read your answers before the call, so we can get straight to specifics.
We use your answers only to prepare for the call.
Step 1 of 6
How many people work in your company?
Questions before we start.
Who is the AI director at CloudTech?
Someone from our team with experience in system design and AI delivery. They own the subject on your side: set priorities, prepare change specifications and coordinate delivery.
How much time does it take from our team?
At the start, a few conversations with people who know the processes. After that, a monthly meeting and one person who signs off changes on your side are usually enough.
Does our data need to be in order first?
No. Assessing the data is part of the work. If it needs cleaning up, we plan for that before proposing anything built on AI.
Will our data be sent to external models?
Only with your consent. For every change we agree where data is processed. Where needed, we use models running in your own cloud or infrastructure.
Can we work with our current IT provider?
Yes. The AI director can prepare the specification and sign off work delivered by another provider or your own team.
What if no change is worth making?
Then we will say so. Sometimes the best recommendation is a simpler change to a process or an existing system, without AI.
How do you price the work?
The review is free. The delivery map is priced as a project and the AI director as a monthly fee that depends on scope. Each delivery is priced separately, from its specification.
The first step is a conversation with an engineer.
The free review takes a few minutes of your time. Then we decide together whether there is a good place to start.