Skip to content
Intlexl

PM agents

Six agents. You set the limits.

Agents do the follow-up, the reporting and the watching. You choose how far each one may go, and everything it produces says whether a rule or a model wrote it.

Book a demoRead about governance
Agents
6
Autonomy tiers
3
Observe, Propose, Act
Marks on every output
2
AI or Automated
Views on the Agents page
5
For you, Program, Agent configuration, Settings, History

Autonomy

You set how far each agent may go.

  1. Observe

    Reads your plans and reports what it finds.

    Findings appear for the right people. Nothing is changed.

  2. Propose

    Drafts the action and waits for a person.

    Each proposal shows what approving it will do, with Approve and Dismiss.

  3. Act

    Carries out approved kinds of action on its own.

    Only where an administrator has enabled it, inside the limits you set.

Agent 01

Can use AI

Status Chaser

Follows up on overdue work for you.

Takes over: Chasing stale tasks by message and meeting.

  • Finds tasks that are overdue and not done
  • Asks each owner in the app, on a cadence you set
  • Records the answer: done, still working, new date or blocked
  • Escalates by policy and keeps a history for every task
Status ChaserCan use AI29 Sept, 21:06

Priya Raman: 3 overdue tasks on Customer Portal. The oldest is Payment gateway failover test, which is 13 days overdue.

A model can word the request. Tasks, dates and counts are always automated.

Agent 02

Automated

Standup and Status Report

The daily report, written before anyone asks.

Takes over: Compiling the daily standup and status update.

  • Compiles a daily standup across every project in scope
  • Produces a status report from the same data
  • Lists what changed, what is late and what is blocked
Standup and Status ReportAutomatedExample of what it writes

Daily project health, 30 September: 9 tasks past their deadline in 4 plans, led by Customer Portal with 6, and no dependency cycles.

Agent 03

Automated

Intake Triage

No request sits without an owner.

Takes over: Checking the intake form for new requests.

  • Checks submissions to your intake forms every day
  • Flags any request nobody has picked up
  • Puts it in front of the people who can assign and prioritise it
Intake TriageAutomatedExample of what it writes

Supplier onboarding portal: a new request with no owner.

Agent 04

Automated

Risk and RAID

Spots the risk and puts it on the register for approval.

Takes over: Keeping the RAID register current by hand.

  • Detects risks from schedule, dependency and workload signals
  • Proposes a register entry for each one, logged once a person approves (straight away at Act)
  • Leaves every entry visible and reviewable
Risk and RAIDAutomatedExample of what it writes

Proposed for the Risk Register: Tomas Bauer is over-allocated at 130% in the weeks of 5, 12 and 19 October.

Agent 05

Automated

Forecasting and Early Warning

See the slip before the date moves.

Takes over: Finding out a plan is late when it is already late.

  • Projects whether each plan will hit its dates
  • Flags the plans that are likely to slip, with how far
  • Shows the overdue and upcoming work behind the forecast
Forecasting and Early WarningAutomated29 Sept, 21:06

Forecast: "Customer Portal" likely ~13 days late. 6 overdue, most overdue Payment gateway failover test.

Agent 06

Can use AI

Resource Profiler

Know who has done what.

Takes over: Asking around for who has the right skills.

  • Builds a directory of people and their expertise
  • Infers it from the tasks they are assigned across your plans
  • Refreshes as the work changes
Resource ProfilerCan use AI29 Sept, 21:06

Create the Resource Directory: 8 people to profile, found from the owners across your plans.

A model can infer expertise tags. Task counts and workspaces are automated.

Decisions

Proposals wait for a person.

At the Propose tier an agent shows what approving will do. Nothing changes until someone approves it, and the decision is on the record.

The Agents page

Everything the agents found, in one place.

FAQ

Questions people ask.

PM agents are six standing routines built into Intlexl that do the coordination work a project manager would otherwise do by hand: chasing status, compiling reports, triaging requests, keeping the risk register, forecasting slips and profiling expertise.

Each agent runs at Observe, Propose or Act, chosen per agent by your organisation. Observe reports only. Propose drafts an action and waits for a person to approve it. Act carries out approved kinds of action on its own, and only where an administrator has enabled it.

Every output is marked. Automated means rules applied to your data produced it. AI means a model wrote that part, for example the wording of a message. Counts, dates and task lists are always automated, and the Status Chaser and Resource Profiler are the only agents that can use a model at all.

On the Agents page. For you lists what needs your answer, Program shows every open finding and the decisions waiting, Agent configuration sets each agent's schedule and scope, and History keeps findings, chases, reports, runs and tool calls, with retention you set.

Agents work inside the permissions of the organisation and the people involved. Recipients of a chase must be active members of the organisation, and every run is recorded in the run log.

Bring one plan. See it run.

In 30 minutes we set up a workspace like yours and show sheets, Gantt, dashboards, capacity and a PM agent at work.

Book a demoExplore the platform