Planograph AI

A plan that accounts for reality

A backlog is only part of planning. There are holidays, specialisms, unfinished work, delivery stages, and dependencies on other products. A spreadsheet can hold any date; delivering against it is harder.

Planograph AI is my own development-planning assistant. It connects backlog structure, people’s availability, and delivery rules, from annual scope estimates to assignments across specialists and sprints.

The questions it helps answer

Management question	Capability
What makes up an initiative?	An epic, feature, story, and task hierarchy
What depends on what?	Dependency graphs and execution order
Who can do the work?	Teams, roles, skills, availability, and capacity constraints
What fits into the period?	Capacity calculation and sprint allocation
How does an annual intent become delivery work?	Annual estimates, delivery breakdown, and transfer of an agreed plan into the workflow
Where is the plan at risk?	Conflicts, overload, dependencies, and deviations
How can we discuss it elsewhere?	A visual workspace and Excel exports

Mathematical optimisation and AI have different jobs

The allocation engine uses OR-Tools CP-SAT to model resource limits, dependencies, and planning rules. ML components support effort estimation. Planning agents orchestrate defined calculations and transfers; they are not a claim that an LLM invents a delivery schedule. Plan preparation and agent workflows connect inputs with subsequent actions.

A feasible plan respects the supplied constraints but may not be the best possible plan. Reaching a time limit without a solution does not prove that the plan is impossible; CP-SAT distinguishes these outcomes through different statuses. A solution also cannot correct inaccurate effort estimates: the team still needs to check those.

Assumptions remain important: a confident answer from a model cannot establish that a team has enough time. If the work does not fit, the useful result is an explicit conflict and options: change staffing, sequence, scope, or the date.

A practical scenario

Several products share specialists. Analysis precedes development and testing, while some activities can run in parallel. I set availability, roles, dependencies, and priorities, inspect the calculated plan, and compare alternative allocations. If an urgent task arrives, the plan should expose what it displaces.

What a demonstration covers

We can take an illustrative backlog with several initiatives, shared specialists and dependencies, and follow one complete scenario:

1. Build the task hierarchy and define people’s availability.
2. Calculate assignments and see where capacity is insufficient.
3. Change a priority or a specialist’s availability and compare the consequences.
4. Prepare the transfer of an approved annual plan into an operational period.
5. Review the preview and apply precisely the agreed version.

The current annual-planning agent calculates in a disposable data copy: hierarchy transfer, people balancing, release sequencing and capacity checks. Blocking conflicts prevent application. Revision checks ensure that only the reviewed plan is applied to the expected source state; recovery is supported.

For a leader, the value is concrete: see the consequences before changing the working plan. A conflict should remain visible instead of becoming an impossible date behind a reassuring status.

Why “Planograph”

This is a working name for planning through connections between tasks, people and deliveries. AI can support preparation and analysis; mathematical constraints and human acceptance retain distinct roles. The name can evolve without changing that principle.

What adoption requires

We begin by mapping your process: backlog structure, effort estimation, roles, availability and dependencies. Then we test the calculation on representative tasks and compare it with your current planning.

Adoption in another company requires adapting data, permissions, integrations and operations. Each connection needs its own readiness check. The current stage is my own working product, with preparation for external deployments.

Demonstration, adaptation or acquisition

I am open to a demonstration, implementation in your organisation, joint development, or acquisition of the product and its underlying work. Delivery scope, rights, support, and the roadmap are agreed for each engagement.

Discuss Planograph AI →