Skip to main content

Councils & government

AI software for councils and government.

We design and build agentic software that helps public-sector teams gather information, prepare work and carry out authorised tasks. People retain control of consequential decisions, with clear permissions and a record of the work performed.

Support the work behind the service.

In local government and across the public sector, a service request can cross several teams, records and systems before anyone has enough context to act. Officers need to understand what has happened, what is missing and who has authority to take the next step.

We build shared workspaces around those needs. People can inspect information and work directly, while AI agents take on suitable tasks within defined limits.

Gather the relevant context

An agent could bring together records it is authorised to access, identify missing information and prepare a brief for an officer to check.

Prepare work for review

It could organise supporting material or draft a response using the available context. A responsible person reviews the work before an action that requires approval proceeds.

Coordinate the next step

It could flag an outstanding task, route work to the appropriate role or identify an exception that needs attention, within the agreed workflow rules.

These are examples of work to assess during discovery. The appropriate scope depends on your systems, information, policies and decision responsibilities.

Authority belongs in the software.

We design the software so people and agents act through the same application rules and permission checks. An agent's role determines what it can see and do; the consequences of an action inform the review it requires.

Defined permissions

Specify which records and actions each role can access. Apply those checks when an action is requested, whether the request comes from a person or an agent.

Human review and intervention

Identify where approval is required, who provides it and how work is escalated. Make active work visible so authorised people can intervene and handle exceptions.

Records that explain the work

Define the record needed for each action: who or what acted, under whose authority, with which approvals and with what result. Agree access and retention requirements for those records.

Controlled failure and recovery

Plan what happens when information is incomplete, a system is unavailable or a task cannot be completed. Establish which changes can be reversed and which need approval before they happen.

Make assurance part of discovery.

Before a pilot or build, we work through the operating requirements with the people responsible for the service, technology, information and assurance.

Information and hosting

Establish what data the solution may use, where processing would happen, what third parties are involved and the applicable access, retention and hosting requirements.

Accessibility and daily use

Understand how staff will inspect information, make decisions and complete tasks. Define accessibility requirements and test the interface with the intended users and working conditions.

Integration and procurement

Examine existing systems, available connection points and supplier constraints. Identify the procurement steps and evidence needed to assess the proposed engagement.

Evaluation and ongoing ownership

Agree how to test permissions, approvals, exceptions and outcomes. Identify who will operate the service, review its performance and authorise changes after launch.

The required controls, assurance evidence and acceptance criteria are agreed for each solution and engagement. Plumb Digital's ISO 27001 certification is in progress.

Start with a bounded piece of work.

  1. 01

    Define the service outcome

    Choose a workflow with a clear owner and a problem worth addressing. Map the people, decisions, systems and information involved.

  2. 02

    Agree the limits and evidence

    Define what the software may do, what people must review and how success will be assessed. Resolve critical dependencies before expanding the scope.

  3. 03

    Evaluate before extending

    Test the proposed capability against realistic cases, including exceptions. Use the findings to decide what is ready for use, what needs to change and whether further investment is justified.

In development: vegetativeAI.

vegetativeAI is our product in development for council vegetation enforcement. It is being designed to guide teams from the first report through evidence gathering and case preparation, with AI assistance that a person reviews.

It brings our approach into a specific council workflow: keep the context together, make the next step clear and preserve human judgement throughout the case.

Questions from public-sector teams.

Yes. Discovery can establish the intended outcome, delivery options, operating constraints and evidence needed for a decision. Its scope and outputs are agreed around the questions your organisation needs answered.

Bring the workflow into the conversation.

Tell us which service or internal process you want to improve, who is responsible for it and the constraints that matter. We can work through the opportunity and an appropriate starting scope.