[ 01 ]
We build it, not just plan it
The audit ends in a ranked list and a scoped pilot. The pilot ends in something running in your environment, not in a slide deck.
[ AI CONSULTING & AGENTIC WORKFLOWS ]
AI, put to work.
You have tried the chatbots. niivo finds the workflow worth automating, builds the agent in your own environment and measures it against a baseline. After 4 weeks you see the numbers and decide: go or stop.
Fixed-price audit: CHF 4,900Go or stop after the pilotYou own what we build
[ 01 — SOUND FAMILIAR? ]
The tools are the easy part.
- We bought Copilot licences. Three months later, hardly anyone uses them.
- Order intake takes our team two days a week, retyping emails into the ERP.
- Every inbox and ticket queue is sorted by hand, and the rules live in one person's head.
- Everyone has tried a chatbot. Nobody knows what is allowed with company data.
- Every AI idea stalls in a review that nobody prepared for.
niivo starts with the work, not the tool. We measure the process, pick where AI pays off first and agree the number to beat before anything is built. IT, security and compliance are in the room from day 1, not at the end.
[ 02 — WHY NIIVO ]
Six reasons teams work with us.
Consultancies deliver decks. SaaS vendors deliver a chatbot for the average company. We deliver a working workflow for yours.
- [ 02 ]
Vendor-neutral by design
OpenAI, Anthropic, Azure AI Foundry or open-source models, chosen per workflow on quality, cost and where the data may go. When a better model arrives, we re-test it on the same sample before you switch. - [ 03 ]
Your data stays where you decide
Your tenant, a cloud region you choose or your own servers. We document which provider sees what, and local models handle the cases that cannot leave. - [ 04 ]
Every agent is accountable
A named owner on your side, visible sources and prompts, a log of every step and an off switch that does not need us. - [ 05 ]
Measured, with a decision
Each pilot has a baseline and a number to beat, agreed in writing. After 4 weeks you choose to go, extend or stop, and running costs are stated before you do. - [ 06 ]
Small and fast, with operations behind it
niivo is the AI brand of worxspace. Its ops team runs identity, security and tenant operations, so your agent does not depend on two people and a laptop.
[ 03 — THE QUESTIONS WE HEAR FIRST ]
Honest answers to the hard questions.
These come up in almost every first conversation. We would rather answer them here than in the third meeting.
Will our data leave the company or end up training someone's model?
Only if you decide so, workflow by workflow. We classify the data first: what may go to a provider, what stays in your cloud region, what stays inside. Business API terms of the main providers generally exclude training on your data, but terms differ and change, so we read them with you and write down the result. Data that cannot leave is handled by local models.
AI makes things up. How can we trust the output?
You should not trust it blindly, and the workflow is built so you do not have to. Agents show their sources, stop and hand over when unsure, and nothing is sent, paid or deleted without a person's approval. The pilot report counts every correction, so you see the error rate before you scale.
We tried ChatGPT or Copilot and it did not stick.
Tools stick when they are tied to a task. We start from a workflow, or from three daily tasks per team, show them once in your own documents and check usage after 30 days. Licences that nobody uses get flagged, not renewed.
IT, security or compliance will block it.
They are invited on day 1 and get what they need to say yes: a data flow map, access rights, an identity of its own for each agent, logs and an off switch. Identity and tenant security are run by worxspace's ops team, so these answers are practice, not theory.
Pilots never end and the return is unclear.
Ours end after 4 weeks with a decision. The baseline is measured before we build, the number to beat is agreed in writing and running costs are stated up front. If the pilot does not beat the baseline, we stop it.
Models change every month. What about lock-in?
We are vendor-neutral and keep the workflow separate from the model. Prompts, code and the test sample are yours, so a new model is tested against the same sample before you switch. There is no proprietary platform to move out of.
Our team will resist, or fear being replaced.
The agent takes over the retyping, not the judgement. The people who run the workflow are interviewed first, so it is built around how they work, and they check and approve the output. The team stops retyping orders. They still check every one before it goes out.
Who runs it once you leave?
You do, or we do. The handover includes documentation, a runbook and a named owner on your side, and the off switch works without us. If you want support, that is a separate agreement backed by worxspace's ops team.
How is niivo different from a consultancy or a SaaS vendor?
A consultancy hands you a report. A SaaS vendor hands you a product built for the average customer. We hand you a working workflow in your environment, measured against your baseline, with the documentation to run it without us.
What about revDSG, GDPR and the EU AI Act?
We build the controls: masking before prompts, data residency decisions, audit trails and an inventory of AI uses. Legal interpretation stays with your counsel, and we do not promise compliance. We prepare the material for your compliance team to assess.
[ 04 — SERVICES ]
Eight ways we put AI to work.
Every engagement starts with the audit. From there you pick what fits the workflow: an agent, a phone line, a local model, a classifier, Copilot or a training course.
- [ 01 — AUDIT ]
AI workflow audit
Fixed price, 2 weeks. You get a baseline, a ranked list and one pilot recommendation, and the report is yours whether or not we build. - [ 02 — AGENTS ]
Agentic workflows
Stop retyping orders, invoices and tickets. The agent drafts, a person approves, every step is logged and there is an off switch. - [ 03 — VOICE ]
Voice agents
No more missed calls. Phone assistants take intake, bookings and first-level questions, and hand over to your team with a written summary. - [ 04 — SOVEREIGN ]
Sovereign & local AI
Use AI on data that cannot leave. You choose where data lives and where the model runs, from your cloud region to your own servers. - [ 05 — CLASSIFICATION ]
Classification at scale
Sort thousands of documents, emails and tickets with compact models, run locally or via an API, with accuracy measured on your own sample. - [ 06 — COPILOT ]
Copilot rollout
Turn licences into a habit. Readiness check, a pilot group and three tasks per team, with usage measured after 30 days. - [ 07 — REGULATED ]
AI in regulated environments
Get past the compliance review with controls you can show: masking before prompts, data residency, audit trails and a usable policy. - [ 08 — ACADEMY ]
Academy
Practical courses for teams, leads and compliance, taught by people who build. AI basics, prompting, agents, governance and Copilot.
[ 05 — THE PATH ]
Audit, pilot, rollout, measure.
Four steps. Each one ends with something in your hands and a decision that is yours. No step starts without the one before it.
- 01
Audit
We map your workflows, measure a baseline for each and rank them by hours saved.2 weeksYou getA ranked list, a baseline per workflow, data-risk notes and one pilot recommendation.You decideWhich workflow, if any, is worth a pilot. - 02
Pilot
One workflow, one agent, a number to beat. The agent drafts, a person approves every case.4 weeksYou getA weekly update: volume, corrections and time per case.You decideWhether to adjust the scope while it runs. - 03
Rollout
Only after a go decision. Monitoring, audit trail and a named owner move into production.After a go decisionYou getA running workflow with documentation, a runbook and stated running costs.You decideGo, extend or stop, after comparing the pilot with the baseline. - 04
Measure
A monthly review of what the agent did, against the baseline.OngoingYou getA monthly report with volume, corrections and time saved.You decideKeep it, extend it or stop it.
[ 06 — RISK-FREE START ]
The risk stays on our side of the table.
These are commitments, not results.
4,900
CHF · FIXED-PRICE AUDIT
Scope and price are agreed before we start. No open hours, and the report is yours whether or not we build.
4
WEEKS TO GO OR STOP
Every pilot ends with a decision against the baseline. If it does not pay off, we stop it.
100%
YOURS: PROMPTS, CODE, DOCS
You own what we build. Prompts, code, configuration and documentation are handed over, with no proprietary platform to leave behind.
1
OFF SWITCH, NO CALL TO US
Every workflow has a named owner on your side and an off switch that works without us.
[ 07 — WHO IT IS FOR ]
Built for the people who have to make it work.
niivo fits when someone in the company owns a problem and needs it solved, not demonstrated.
- [ 01 ]
COO or head of operations
You own a process that eats hours every week. You want a number before you commit budget, and a plan for what happens after the pilot. - [ 02 ]
IT lead
You want AI in your tenant with identities, permissions and logs you can defend, instead of another shadow tool that nobody manages. - [ 03 ]
Compliance and data protection
You need to see where data goes, what is logged and who owns each use before you can say yes. We prepare that material for you. - [ 04 ]
Team leads
Your team retypes, sorts and chases. You want the routine gone and the judgement kept, and a say in how it is built.
Typical situations
- Mid-size companies with a Microsoft 365 tenant and a backlog of routine work
- Finance, insurance, health, legal and public-sector teams with strict data rules
- Teams that bought Copilot licences and need adoption, not another launch email
- Operations with shared inboxes, order intake, invoices or a busy phone line
- Companies that want AI without sending sensitive data to a public model
[ 08 — WHAT A PILOT REPORT LOOKS LIKE ]
A number every week, not a slide at the end.
This is the format of the weekly pilot update. You see volume, corrections and time per case while the pilot runs, so the decision is never a surprise. The figures below are made up to show the layout.
PILOT — WEEK 3 OF 4 — ORDER INTAKE
212
Orders drafted by the agent
198
Approved unchanged
9 → 2
Minutes per order
14 orders were corrected by the team. Next: add the supplier catalogue. Rollout decision in week 4.
[ 09 — PLATFORMS ]
We pick the model for the job.
niivo is not tied to one vendor, and nothing we build is either. We choose per workflow, based on quality, cost and where your data may go.
- OpenAIFrontier models via API
- AnthropicClaude models via API
- Microsoft Azure AI FoundryModels in your Azure tenant and region
- Open sourceOpen-weight models and projects, run locally or hosted
[ 10 — WE ARE NOT ]
What niivo is not.
- An AI hype agency.
- A chatbot vendor.
- A strategy deck with nothing built.
- A black box you cannot switch off.
- A replacement for your team.
niivo is the AI brand of worxspace, which runs Microsoft 365 and enterprise IT for its clients. We are small and fast, and an ops team stands behind us: where a workflow touches your tenant, worxspace runs and secures identity, permissions and operations, with a written handover between us.
Bring one workflow that eats your time.
30 minutes, plain language, no pitch. We sketch it together and tell you honestly whether AI pays off. If it does not, you have lost half an hour.