The ever changing reality in recent years is pushing business to rethink their entire business planning paradigm. Economic instability, disrupted supply chains, and the rapid growth of data volumes are driving increased interest in AI/ML-powered business planning platforms. And this interest is well justified. Intelligent platforms offer better forecasting and service levels, reduce supply chain management costs, lower reliance on human expertise, and, consequently, decrease labor intensity. These platforms can handle tasks like forecasting, optimization, and scenario planning more efficiently, freeing employees from routine operations.

For many market players, implementing AI/ML-based Integrated Business Planning (IBP) platforms is a brand-new experience. Yet, there are already companies successfully using them in practice. Newcomers are often intimidated by the cost, although contrary to popular belief, intelligent platforms are not more expensive than traditional ones—and they deliver far more business value. Moreover, project costs can be significantly optimized if the process is properly structured from the beginning. The more factors you account for during the analysis and design phases, the fewer costly revisions you’ll need during implementation.

How can you minimize your starting budget and prepare for a project to ensure it stays within planned timings and cost?

1. Structure your vision and expectations of the target business process

Take an end-to-end business process like promo planning, which involves multiple departments—planners, trade marketing managers, finance, and IT. Each participates to varying degrees, but often with no shared KPIs or reporting lines. While the process may function under normal conditions, no one truly has a holistic view, let alone proper documentation.

Before implementation begins, the existing process needs to be documented and evaluated. Decide whether you’re automating it “as-is” with minimal tweaks or whether it requires full reengineering. Process reengineering is often inevitable in launching IT-platforms, but it does impact timelines. Clearly understanding the process to be automated will make life easier for developers, speed up implementation, and reduce final costs.

2. Clean and organize your data

Timely auditing of all relevant data, including master data that will feed into the planning platform, is critical to cost control. In one project, we had to load shelf sales data from key retail chains to identify correlations between manufacturer shipments and point-of-sale sales using ML algorithms. However, the data came in multiple formats—some in Excel, others from retailer portals, etc.

Instead of moving on to solving the core task, we had to work with the client to unify data formats—in essence, restructuring their internal process. This kind of preparation is best done before feeding data into the system. Otherwise, forecasts must be manually loaded each time, increasing not only implementation costs but also total cost of ownership—including licensing and support. If your data is unstructured, inconsistent, or contains errors, address these issues before the project starts to avoid unnecessary expenses.

3. Prioritize the most impactful data

IBP platforms—both traditional and AI-powered—rely on large volumes of input data. The more modules and logic you want to include, the more data you’ll need. Some clients attempt to load as much data as possible—historical sales, promo calendars, price fluctuations, weather, inventory levels, ad spend, macroeconomic indicators, etc. While possible, this approach makes the project more expensive, complex, and time-consuming—and not all of these data sources will improve forecast accuracy and other business critical KPIs.

In one case, a client opted for a highly complex planning logic for raw material optimization, involving dozens of constraints—line capacities, inter-site transfers, and more. The extensive input data significantly raised both the timeline and cost, though the company could have benefited from a simpler version.

We typically recommend focusing on key data sources that address core business needs during the initial launch. Additional constraints and planning horizons can be added later as the platform evolves.

4. Stick to the original project scope

The project scope is defined during the analysis and design phases—what functionality is needed, which business process is being automated, etc. Based on this, the vendor and integration partner allocates resources and estimates cost. However, during implementation, clients often identify “nice-to-have” features not included in the original scope.

While some additions are reasonable, others can lead to scope creep, increasing time and cost without significantly improving results. In one project, we spent considerable time designing complex workflow logic, only for the client’s team to use just a portion of it after launch—which was sufficient for daily operations.

To control the budget, it’s best to deviate from the initial scope only when absolutely necessary.

5. Ensure collaboration at all organizational levels

Implementing complex solutions—especially AI-based ones—requires deep engagement from internal teams: process owners, project managers, and IT. While having an executive sponsor is important, it’s not enough. In one case, the IT team was not supportive simply because no one had explained the project’s purpose or value to them.

Team motivation is the project manager’s responsibility. But they also need to manage stakeholder expectations, maintain clear communication between client and vendor, ensure quality and deadlines, and keep everyone aligned. If they lack experience managing large transformation projects or aren’t committed to the outcome, staying within budget and schedule becomes nearly impossible.

6. Avoid parallel implementation of multiple projects

Running several implementation tracks at once overloads the team, reduces efficiency, and drastically increases costs. If time allows, it’s better to roll out platform modules (e.g. Demand Planning, Supply planning, FP&A etc.)  sequentially.

Avoid starting transformations of connected systems (like ERP) during IBP implementation. One client decided to change their database midway through our project, which required us to rewrite platform architecture just before the launch. As a result, the project was delayed and went over budget.

Running multiple transformation projects in parallel raises the risk of failure across the board—not to mention increased costs, timelines, and critical staff overload.

7. Avoid rigid deadlines

In recent years, we often hear, “We need it yesterday.” Urgency is rarely budget-friendly. But if a system is about to fail, there may be no choice but to spend more to keep the business running. Deadlines may also be driven by business cycles or corporate transformation roadmaps, where a delay in one project disrupts the entire plan.

These constraints are often non-negotiable—but they introduce risks that tend to materialize. Almost inevitably, deadlines shift anyway.

Regardless of the reason for a hard deadline, both client and vendor must realistically assess feasibility. Even with a larger team working around the clock, some timelines are simply unachievable. All parties need to weigh the risks and decide accordingly.