Consulting for a faster, simpler public sector

What we automated

We automated checks on invoices, purchase orders and incoming stock. The checks compare the purchase documents against the goods received.

Checking the amounts

The checks compare the amounts on invoices, purchase orders and payments to verify that the numbers match. The team can review each transaction with the information brought together.

Payments made correctly and on time

The checks also verify that payments were made correctly and on time. Each payment date and amount is compared with what was due.

What was ordered and what arrived

The checks verify that the product ordered matches what arrived. Differences are visible to the team for review.

Scattered information

Information came through different channels and in different formats. There were paper documents, emails, WhatsApp messages and records held only in physical office files. These often referred to the same client, supplier or project.

One intake process

We changed the intake process and added automations to bring those sources into one flow. Information no longer stays separate just because it arrived through a different channel.

Connecting the records

We linked records that referred to the same client, supplier or project. Information that used to be held separately is now available through one source of truth.

Information for decisions

Senior management can see the information related to each client, supplier or project more clearly. That view helps them make better-informed decisions faster.

Practical training

We worked with middle and senior managers on using AI tools. The training was practical and focused on how they could apply those tools to their own tasks.

Individual support

We helped each person examine their own workflow and find ways to improve it with AI. The individual sessions helped them apply what they had learned to the processes they manage.

After the training

After the more intensive phase, we held follow-up sessions. We used those meetings to answer questions and keep helping with the new processes they had started using.

The team's main AI tool

We built a shared tool that the team uses as its main AI tool. It connects to context about the company and each client. Each person can consult the information they have access to.

Conversations and purchase orders

The team can look up what was discussed with a client or check the status of a purchase order. They can find that information through the same tool, without searching for it in separate places.

Cross-referencing information

The tool supports simple cross-checks between available information. For example, the team can connect what was agreed with a client to the status of a related purchase.

Context that updates automatically

Automations collect new information and update the shared context. The team can ask about a process or client without having to rebuild that context manually for every query.

What we did

We made all of a company's processes accessible, auditable and operable through AI agents. The team can review the information the agents use and what happens in each process.

How we organized the information

All information coming into and going out of the company is collected and stored in a format the agents can read and use. They can then find the context they need to carry out each process.

What AI-native means

AI can work with the company's processes and information without a person having to gather and explain them again for every task. The information is available for both people and agents to use.

The sale of the company

The company was sold later. In that transaction, being recognized as AI-native contributed to a higher valuation multiple and a higher sale price.

Preparing the purchase

We made the information people need to propose a purchase visible to the team. The agent brings that information together and compares supplier quotes so the team can review the request.

The team makes the decision

The team reviews the proposal and can approve or edit it. Only after approval does the agent create the purchase order.

Checking what arrives

Once the purchase has been made, the agent matches delivery documents against the orders. It checks that what arrived matches what was requested and flags differences for the team to review.

What we automated

The agent covers much of the process, from preparing the request to checking the delivery. People keep control over what to buy and on what terms.