Sales route planning with AI: how to optimize visits and frequency

A sales manager at a cleaning products distributor told me, half resigned, a scene he knew by heart. One of his reps drove more than 200 kilometres to visit a customer who bought once a year, out of habit, because "he was always his customer". Meanwhile, a AAA customer located ten blocks from the office saw him every two months, when they remembered. Visit frequency followed no logic of value: it followed inertia.

That scene repeats in almost every sales force that grew without a system. The routes were built years ago, the reps inherited them, and nobody sat down to ask whether visiting time is being invested where it yields most. The result is money thrown at kilometres and, worse, valuable customers left unattended that a competitor is waiting to take.

The underlying problem is that visit frequency gets assigned by habit, not by value. And the rep's time is the most expensive and scarcest resource the commercial area has. Every hour on the wrong route is an hour not spent where it could sell more.

In this article you will see why inherited routes cost more than they seem to, how a layer of AI on top of your ERP builds the optimal frequency and itinerary per customer, a real case of a distributor that redesigned its routes, and the steps for doing it without chaos.

---

1. The problem: routes inherited, not planned

1.1 Frequency by habit

At most distributors, the question "how often do we visit this customer?" was never answered with data. The frequency was set years ago and froze there. There are customers receiving weekly visits out of inertia even though their consumption fell, and customers who grew but still receive the attention they got when they were small.

1.2 The hidden cost of a badly invested kilometre

The cost of a badly built route is twofold and almost always invisible:

• The direct cost: fuel, time and wear on visits that generate no return.
• The opportunity cost: the high-value customer who gets neglected and cools off, or who leaves outright for the competition because they stopped feeling close to you.

The second is far more expensive than the first, and it is the one nobody sees until the customer is already gone.

1.3 Why the spreadsheet does not solve it

Building routes in a spreadsheet is a puzzle impossible to optimize by hand: you have to cross each customer's value, potential, geographic location, the rep's capacity and time constraints. A person can produce a reasonable version, but not the optimal one, and above all cannot recalculate it every time something changes.

2. What AI does with route planning

2.1 Assigning frequency by value, not by inertia

The first change is conceptual: visit frequency is defined by each customer's value and potential, not by habit. The AI layer crosses how much they buy, how often, what potential they have and what happens when visits stop, in order to assign the frequency that maximizes return. This rests on the same logic as RFM segmentation: not all customers are worth the same, and attention should reflect that.

2.2 Optimizing the geographic itinerary

With frequency defined, AI builds the itinerary: it groups visits by area, minimizes dead kilometres and respects each rep's real capacity (how many quality visits fit in a day). The goal is not to squeeze in more visits, it is to fit the right visits with the least waste.

2.3 Recalculating when something changes

What a spreadsheet cannot do, AI can: recalculate. If a customer raises their consumption, if a new one comes in, if another cools off, the frequency and the route adjust. Planning stops being a dead document and becomes something living, connected to your ERP data, without replacing the system.

3. A real case: a distributor that redesigned its routes

3.1 Before

A beverage distributor with 14 reps and several hundred active customers had routes built more than five years earlier. Some reps did enormously long itineraries with many low-value visits, while important customers in central areas received less attention than they warranted. Nobody had a clear view of whether the effort was well distributed.

3.2 A phased implementation

1. Diagnosis (month 1). Sales, current frequency and the location of each customer were crossed to see where time was badly invested.
2. Frequency redesign (month 2). Frequency was reassigned by value and potential, raising attention to AAA customers and lowering it on low-return accounts.
3. Itinerary optimization (month 3). Geographic routes were rebuilt with the new frequencies, reducing dead kilometres.

3.3 After

The redesign freed a significant share of the reps' time, on the order of 15% to 25%, that used to go into travel and low-value visits. That time was reinvested in the highest-potential customers and in prospecting. AAA customers started receiving the attention they deserved and several accounts that had been cooling off were reactivated. The sales force did not grow in size, but it produced considerably more with the same people.

4. Step-by-step implementation

4.1 Get the customer data in order

Before optimizing, you need to know how much each customer is worth and where they are. That already lives in your ERP: sales, purchase frequency and address. The first step is organizing it.

4.2 Define the target frequency by segment

With customers segmented by value and potential, you define how much attention each group deserves. A AAA customer with room to grow is not visited like a marginal, stable one. Connecting this to a good AI strategy for sales ensures the frequency serves concrete commercial objectives.

4.3 Build the itineraries and bring the reps in

AI proposes the itineraries, but adoption is won with the team. It pays to involve the reps, who know details of the territory no data captures, and to present the redesign as a tool for selling more rather than as surveillance.

4.4 Connect it to coverage

Route planning is the foundation for something bigger: increasing market penetration. Once the routes are optimized, the next step is going after the prospective customers who do not buy from you yet, which is what sales coverage and customer penetration is about.

5. ROI and measurable benefits

5.1 What to measure

The key indicators:

• Time on the road versus effective selling time.
• Kilometres per visit.
• AAA customer coverage (what percentage receives the target frequency).
• Sales per rep and per hour of visiting.

5.2 The typical return

The return comes from two sides: direct savings in travel costs and, above all, more sales from reinvesting the freed-up time in high-value customers. When the most expensive resource (the rep's time) is well focused, the same sales force produces more without adding people.

5.3 The benefit for the team

There is a benefit the rep appreciates: shorter, more sensible routes, fewer hours lost on the road and more time in front of customers who genuinely buy. A well-built route also improves the team's quality of life, and that shows up in turnover.

6. Common mistakes in planning routes

6.1 Assigning frequency by how old the relationship is

The most widespread mistake is visiting the "lifelong" customer more often just because that is how it has always been. The age of the relationship is not the same as current value. A customer who bought a lot ten years ago and buys little today should not keep receiving the old frequency. Visits are assigned by present value and potential, not by history or affection.

6.2 Optimizing kilometres and forgetting value

The opposite mistake, when people try to correct course, is falling into pure geographic efficiency: building the shortest possible route without looking at who is being visited. A very short route packed with low-value visits is efficient on fuel and terrible on return. First you decide who and how often (by value), and only then do you optimize the how (the geography).

6.3 Treating the route as something fixed

Routes built once and then frozen age badly. Customers change: some grow, others cool off, new ones arrive. Planning that is never recalculated becomes, within a few months, the same old snapshot you wanted to correct. Frequency and itinerary have to adjust when the data changes, and that is exactly what a layer of AI does and a spreadsheet does not.

6.4 Imposing routes without listening to the rep

Data says a lot, but not everything. The rep knows details of the territory no system captures: that this customer is best visited early, that in that area traffic is impossible at a certain hour, that that account is about to grow. Imposing an optimized route without that knowledge generates resistance and leaves value on the table. The best route combines the calculation with the judgement of whoever walks it.

Ready to have your reps where they yield most?

Inherited routes are one of the most expensive and silent productivity leaks in a sales force. Planning frequency by value and optimizing itineraries with a layer of AI on top of your ERP frees up the scarcest resource you have — your reps' time — and puts it where it generates return.

Want to see how it works in practice? Book a demo and we will show you how to redesign your routes based on each customer's real value.