Skip to main content
Customer story

Jan Zandbergen Group AI assisted production planning with Ask Titan

Planners use Ask Titan in Microsoft Teams on top of the Titan data platform to create a realistic first plan, run scenario checks in seconds, and handle exceptions with more confidence during busy weeks.

How we built it
Meat and protein Planning and production Ask Titan in daily use
Microsoft Teams
AI for Microsoft Teams
Which orders are at risk if line 2 has a technical issue for two hours this afternoon?

Orders at risk (today)

At risk: 4
Retail: 3
Food service: 1

Explanation: checked confirmed orders, line capacity, planned runs and changeover constraints. You can expand each step.

Orders Capacity Rules
If we move the pulled pork run to line 3 in the evening shift, what changes?

Impact summary

  • All four at risk orders remain within service
  • Line 3 load increases, but stays within the evening capacity window
  • Overtime stays within the standard for this week

Note: assumptions include standard changeover time and current staffing. Ask for alternatives if constraints change.

Planners can test scenarios quickly, then confirm the final plan with the team.

Gerwin Hoogendoorn | Manager Information Management @ Jan Zandbergen Group BV

“Titan and Ask Titan has improved our planning process, acting as a helpful assistant that provides clarity and speed while ensuring quality through human validation.”

Context

AI-assisted production planning in a complex food manufacturing environment

Jan Zandbergen Group runs a broad assortment with multiple production steps and a mix of retail, food service, and industry customers. Daily planning requires fast trade-offs between service level, shelf life, yield, and line capacity. The goal was a practical AI assistant for production planning that helps planners move faster without losing ownership of the final plan.

Scope

AI assisted production planning and stock insight, focused on daily planning decisions and exception handling.

Users

Planners, operations and management aligned on shared definitions for orders, stock, capacity and planning constraints.

Foundation

Titan unifies orders, stock, capacity and business rules into one governed model, so questions map to one source of truth.

Outcome

Faster planning cycles and calmer weeks through instant scenario testing, clear suggestions and explanations planners can trust.

Challenge

Many constraints and little time to plan

Planners must combine sales orders, forecasts, stock positions, shelf life targets, yields, and line capacity. Critical context is often spread across systems, spreadsheets, and expert knowledge. That makes planning slower, increases manual checks, and makes scenario planning difficult when priorities shift during the week.

Limited planning window

Time goes into gathering orders, stock and constraints before planning can even start.

Many manual checks

Feasibility checks across tools take effort and are error sensitive, especially during changeovers.

Pressure in busy weeks

Peaks and promotions increase change risk when insight is fragmented and scenarios take too long to test.

Solution

A planning copilot in Microsoft Teams for daily production decisions

Food For Analytics implemented a planning copilot using the Titan data platform as the governed foundation and Ask Titan as the AI assistant. Titan unifies orders, stock, capacity, and planning rules into one model. Ask Titan enables planners to ask questions in normal language, generate realistic production suggestions and get clear explanations that support human validation.

  • One governed model for orders, stock, capacity and planning rules
  • AI suggestions that check feasibility against constraints and shelf life targets
  • Clear explanations, so planners can trust, adjust and stay in control
How planners used it

A simple daily loop with proposals, checks and scenarios.

Start

Planners open one view with live orders, stock and capacity signals.

Propose

Ask Titan suggests a realistic plan based on rules and current constraints.

Check

Feasibility checks flag risks on shelf life, yield, allergens, packing or line load.

Adjust

Planners test scenarios in seconds and commit the final plan with full context.

Explanations and checks built in
Governed model under the hood

Results

Results observed in daily production planning

he biggest change is speed and confidence. Planners spend less time collecting inputs and running manual checks, and more time handling exceptions and making decisions. Scenario questions are answered faster, which makes planning meetings calmer and more focused on trade-offs.

Planning speed

-30%

time needed to build the first plan

Planners start with a realistic draft plan and spend time on exceptions, not on collecting and stitching inputs.

Draft plan first
Exceptions focus

Validation effort

-40%

Fewer manual checks

Feasibility checks across orders, stock, shelf life and capacity are handled faster with clear explanations.

Rule based checks
Clear explanations

Adoption

Daily

Ask Titan used in planning meetings

Scenario questions are answered in seconds, so meetings focus on trade-offs and decisions instead of validation.

Fast scenarios
Meeting ready

What we built

A foundation for AI-assisted production planning

This was not a one-off demo. We delivered a reusable set of building blocks: a governed planning model in Titan, a rules and feasibility layer that reflects how planners work, and Ask Titan workflows in Teams with guardrails and explanations. This foundation can be extended to more product groups, planning questions, and teams without rebuilding the core.

Governed planning model

One shared layer for orders, stock, capacity and master data, so Ask Titan and Power BI use the same definitions.

  • Unified view of orders, forecasts, stock positions and allocations
  • Capacity and load per line, per shift, with clear planning horizons
  • Shared measures for service risk, stock at risk and exception flags

Planning rules and feasibility checks

A rule layer that reflects how the planning team works, so suggestions stay realistic and actionable.

  • Priorities, shelf life rules, minimum runs and changeover constraints
  • Feasibility checks against capacity, cut plan and operational limitations
  • Scenario support for what-if questions during daily planning

Ask Titan workflows and guardrails

Prompt patterns, safety rails and explanations so planners can trust the answer and adjust with confidence.

  • Templates for recurring questions: proposals, checks and exceptions
  • Guardrails that keep answers within agreed data, rules and planning horizons
  • Explanations that highlight assumptions, risks and trade-offs in plain language

Adoption routine

A lightweight cadence that makes the copilot part of daily work, and keeps improving the quality of suggestions.

  • Pilot with a small planner group, then rollout to the wider planning team
  • Weekly feedback loop to add questions, refine rules and improve trust
  • Simple monitoring on usage to prioritize improvements

Implementation

Phased delivery from data foundation to a planning copilot

We delivered value in phases. First we aligned on the decisions to improve and connected the minimum required sources into Titan. Then we validated definitions with planners, added planning rules and feasibility checks, and introduced Ask Titan in Microsoft Teams using real planner questions. Finally we embedded the copilot into daily routines and expanded based on feedback and usage.

Phase 1

Scope and connect

Align on the first planning decisions to improve. Connect orders, stock and capacity inputs into Titan and validate the key numbers with planners and operations.

Phase 2

Build and validate

Translate planning rules into a governed model, add feasibility checks, and integrate Ask Titan for proposals and what-if scenarios. Test with a small pilot group of planners.

Phase 3

Run and improve

Embed Ask Titan into daily planning. Capture feedback, expand to new questions and product groups, and keep the rules and definitions stable as usage grows.

Daily feedback loop New scenarios added

Frequently asked questions

What was the main goal for Jan Zandbergen Group?

Support planners with faster daily production suggestions and scenario checks by using Ask Titan on top of a governed Titan foundation.

What did Food For Analytics deliver?

A Titan data foundation plus Ask Titan in Microsoft Teams so planners can ask questions and receive grounded production suggestions with human validation.

How comparable is this case to our situation?

The best comparison is your planning cadence, the number of constraints you manage daily, and how accessible your data is from ERP and planning tools. If you share those three, we can propose a realistic first planning assistant scope.

Is Ask Titan fully autonomous, or do planners stay in control?

Planners stay in control. Ask Titan provides suggestions and explanations, and the team validates and decides. That is critical for quality and trust.

Are the time savings guaranteed?

No. Outcomes depend on the use case scope, data quality, and adoption. We start with one validated planning flow before expanding to more scenarios and questions.

What is a good first step if we want an AI assistant for planning?

Pick one recurring planning decision, define what a good suggestion looks like, connect the minimum required sources, and validate outputs with planners before scaling.

Want a planning copilot that planners trust and actually use?

In a short call we map your planning constraints, data sources and the first daily use case. You leave with a clear rollout that delivers a live planning flow fast and improves through real usage.

Practical session. Concrete plants, rules, constraints and decision cadence.