Omnichannel CPG: What AI Colleagues Would Do in Supply Chain Planning: Issue 07

Obviously, we never want to replace the experts on our team. But as we scale, do we really want to hire more people for every $10M step – or every new warehouse, new supplier, new SKU?

Well yes we do - but only if we don’t know how else to solve it.

Let’s take a step back and define: What are new supply chain hires usually doing at a fast scaling CPG brand?

We’ll look at these three pillars:

1. Demand Planning - Planning demand per sales channel

2. Supply Planning - Translating demand into planned orders and transfers

3. Replenishment - Actually executing those decisions

Now, when does each of these areas get more complex?

📦 A new warehouse: Supply Planning becomes more complex and a new constraint has to be considered for planning (e.g. dynamic safety stock per SKU, capacity constraints & more).

2️⃣ A new channel: Demand Planning gets more tricky - especially if it’s a channel we have not had before like Costco (Retail). In that case, an account manager feeds the plan with their inputs.

🪈 New SKUs:

This has trickle-down effects across all pillars.

→ For demand planning: New forecasts need to be created

→ For supply planning: Decide where the product should live across your network

→ For replenishment: Actually creating those new orders, communicate with the comans and potentially also raw material suppliers

So where’s the biggest opportunity to bring AI colleagues into the team? Where work is both highly technical and highly manual.

Let’s look at both technical work and manual repetitive work:

Technical Work in Demand Planning 
Picking the right forecasting method from a broad range of forecasting methods is hard, especially in spreadsheets.

Technical Work in Supply Planning 
Imagine you’re a CPG brand dealing with constraints like warehouses, inventory policies, safety stock - all the way down to best before requirements per sales channel

…When are we transferring based on those constraints?
…When are we ordering based on them?

How do you bring all of this together? In most cases, spreadsheets just aren’t enough.

Now let’s talk about the manual side of the work – where can an AI employee step in and actually help?

Manual Work in Replenishment

A super overlooked area – but full of exception management requiring manual work.

For example:

Imagine your supplier delays the shipment by 3 weeks due to capacity constraints. This delay was communicated via email.

→ Someone has to open the email or PDF
→ Extract the new ETA
→ Update the planning spreadsheet and ERP
→ Write a response back
→ Lastly, inform the supply planning team to optimize again based on the delay

Now imagine you have multiple warehouses and one runs into a stockout because of this delay.

What are we gonna do?

That’s a complex question requiring complex simulation + manual execution. A great example where an AI colleague could help.

So, what would the perfect AI employee do?

1. Helping with cross-team collaboration by managing the inputs for demand planning.

2. Provide technical support on forecasting and supply optimization

3. Being able to understand all your context when it comes to supplier communication - and help you act on it

4. Being able to address the exceptions which are created through supplier or manufacturing delays - with you being the expert-in-the-loop

More soon.

-Leon