For operations leaders at services firms on Microsoft 365

AI-native software for
complex operations.

Microsoft-native first: the products run on Azure, models route through Azure AI Foundry, and the work shows up in Teams and Outlook. Sign in with your Microsoft or Google work account. Five products in production draft the work from the documents and conversations you already have, mark what an agent wrote, and wait for a person to approve before anything reaches a customer.

Five products are running and open through a conversation.

The operating pattern: prepare a reviewable artifact, obtain the required approval, then execute within that approval. Each product page states which steps and handoffs are available today.
  1. Business signal
  2. Bounded context
  3. Agent preparation
  4. Reviewable artifact
  5. Human approval
  6. Controlled execution
  7. Evidence
  8. Governed outcome

Orange marks where a person decides

  • Entra ID
  • Azure-native
  • Tenant isolated
  • Human reviewed

Five products in production. One way of working.

Each one runs on its own, on Azure, with sign-in through your Entra directory and the work showing up in Teams and Outlook. They are software you license and run, not consulting engagements or custom builds. Each card says where that product stands today, and each product page says what works and what does not, with the date it was checked.

These five are the portfolio as it stands today, not the whole of what is being built. Every product composes with the rest under one identity model and one security posture, and the ones not open yet are listed with their state on the catalogue.

An agent prepares. A person decides. The system acts inside that decision.

Every product follows the same pattern. The agent reads the source it is given, a signed scope of work, a calendar, a call, a ticket, and drafts the record a person would have typed: the plan, the time entry, the deal update, the campaign, the resolution. The draft carries a marker that an agent wrote it, a confidence value and the reasoning behind it, so a reviewer can see why before deciding whether. Nothing reaches a customer, an invoice or a tenant change until that person approves it, and the approval is recorded with the record. The orange marks in the diagram above are those decisions.

Access runs through your Entra tenant, so the conditional access policies you already enforce apply, and the products run on Azure with tenant-isolated storage. Model access routes through Azure AI Foundry by default, which is a default rather than an enforced invariant: a product may call a provider's own endpoint for a specific feature, and each product page says which applies to it.

What works today, with the date it was checked.

Every product page carries a section called "What works today, and what does not". It lists what is built and verified, what is in build, and what is waiting on another company, each with the date the list was last checked against the running system. That section is the proof this site offers, and it is kept current rather than written once.

Every screen and sample on this site uses fictional organisations and fictional data, and says so where it appears. There are no customer logos here yet. When there are, they will be real ones.

Simple Council: evidence your reviewers can examine.

If you are the person who has to sign off on AI agents touching regulated work, Simple Council is the product written for you. It runs a policy engine and a hash-chained evidence ledger for actions routed through its own surfaces, with blueprints from ISO 42001 to CMMC Level 2. Microsoft Purview integration and Agent 365 registry sync are planned, and its page says so, along with what it does not do yet. Its state today: Production, access arranged.

See Simple Council. For what Microsoft itself says about agents in its own products, read About agents in Microsoft 365 Copilot on Microsoft Learn; a web application or Teams tab from us does not establish that a Copilot agent is installable, and each product page states its surfaces.

The questions buyers ask first.

What happens when I request access?
A person reads the request and replies with what the first week looks like for that product. Nothing is created and nothing is charged by the form. You are not signing up for anything.
Do I need all five?
No. Each product is licensed and run on its own. Most firms start with one, usually delivery or time and billing, and add another when the first has earned it.
Where does our data live?
In Azure, in tenant-isolated storage, with the posture stated per product on its page rather than promised across the portfolio. Sign-in is your Entra directory; we do not hold a separate password for your people.
Can the agent act without someone approving?
Not on anything that reaches a customer, an invoice or your tenant. The agent drafts, a person approves, and the approval is kept with the record.
What does it cost?
Pricing is given when access is arranged, because it depends on the seats and the products involved. No price on this site is a price you can buy at today, so none is shown.

What is not on offer today

Stated once, here, so the rest of the page can stay plain. As of 2026-11-08: no product on this site has a self-serve trial or a public price; none is listed on Azure Marketplace yet; no Copilot agent is installable from us; the products run in production for the firms using them, and access for a new firm is arranged with us rather than started from a page. Each product page carries its own list.

Need AI closer to your data?

Simplicity IT Inc., a separate services company, can assess, install and support private AI infrastructure. Simple Intelligence products have their own stated deployment and data postures; this is not a local mode for those products.

Start with the product your firm needs first.

Tell us which one and how your team works today. A person replies.

Prefer to talk first? Write to the team.