Lithuanian grocery retailer IKI is deploying SymphonyAI's Promotional Evaluation and Promotional Planning solutions across its 250-store network, reporting an improvement of more than 10 percentage points in promotional forecast accuracy in initial categories. The implementation highlights how retailers are using vertical AI to predict promotional demand before campaigns launch, connecting marketing decisions more closely with inventory and merchandising operations.

IKI is expanding its use of artificial intelligence in promotional planning, deploying SymphonyAI's Promotional Evaluation and Promotional Planning solutions across its network of approximately 250 stores.

The Lithuanian grocery retailer said its initial work with SymphonyAI across selected categories improved promotional forecast accuracy by more than 10 percentage points and reduced stock-outs during promotional periods. The results were announced by SymphonyAI on September 17, 2026.

For grocery retailers, promotional forecasting is a particularly complex marketing problem. Discounts can change demand rapidly, while the impact of a promotion can vary by store, product, season, competing products and existing purchasing patterns. A forecast that misses demand can create either excess inventory or empty shelves.

SymphonyAI's approach uses AI models to estimate the expected impact of promotions before retailers commit to them. Its promotional optimization technology is designed to project incremental sales, margin impact and potential product cannibalization at product- and store-group levels.

That moves promotional analytics beyond simply measuring whether a campaign worked after it ended.

Instead, the technology is intended to help category managers model different promotional scenarios before launch. Retailers can potentially assess expected lift, margin consequences and interactions between products before deciding which promotion to execute.

For IKI, that capability is being incorporated into a much broader retail operation. The company is Lithuania's second-largest retail chain, according to REWE Group, with around 250 stores and nearly 5,500 employees. IKI has operated in Lithuania since 1992 and is part of Germany-based REWE Group.

The implementation is therefore relevant beyond a single promotional campaign. It gives a large retailer a way to connect marketing planning with category management and operational decision-making.

That integration is becoming increasingly important as retailers attempt to balance price-sensitive shoppers, promotional competition and margin pressure. A promotion is not simply an advertising decision in grocery retail. It can affect purchasing requirements, replenishment, store-level inventory and the availability of products customers expect to find.

AI forecasting can provide a common analytical layer across those decisions.

SymphonyAI says its demand forecasting technology incorporates sales history, promotions, weather, seasonality and local events when generating item-, store- and day-level forecasts. The company reports that other European grocery deployments have produced five- to 10-point forecast-accuracy gains, although those results come from separate customers and should not be treated as equivalent to IKI's reported outcome.

The distinction is important because IKI's more-than-10-point improvement is an early result from selected categories, rather than evidence that the same improvement has been achieved across its entire store network.

The technology also illustrates the convergence of marketing and supply-chain intelligence.

Traditional promotional planning can separate the question of what customers are likely to buy from the question of whether stores will have enough inventory to satisfy that demand. AI-driven planning attempts to connect those variables. Better demand estimates can influence purchasing and replenishment decisions before the promotion reaches shoppers.

The potential business effect extends to waste and stock availability as well. SymphonyAI says its broader retail forecasting deployments have produced reductions in shortages and waste, but these are vendor-reported results from other implementations rather than independently verified results for IKI.

IKI's deployment also provides a potential test case for REWE Group. The retailer's parent operates across 21 European countries, giving successful AI applications within one market a potential pathway to broader evaluation.

The larger MarTech implication is that promotional technology is moving closer to real-time commercial decision-making. Marketing teams no longer have to view promotions solely as campaigns that generate customer response. In modern retail environments, promotional decisions can simultaneously affect demand forecasts, margins, inventory and customer experience.

For IKI, the next stage will be determining whether the initial forecasting gains can be reproduced across additional categories and promotional conditions.

If that expansion produces consistent results, the deployment could become an example of how vertical AI connects promotional strategy with the operational systems required to deliver on that strategy.

Market Landscape

Retail promotional planning sits at the intersection of marketing, merchandising, pricing, inventory and supply-chain management.

AI is increasingly being applied to these functions because retailers have to account for multiple demand signals simultaneously. SymphonyAI's retail forecasting technology, for example, incorporates historical sales, promotions, weather, seasonality and local events into forecasting models.

The market is consequently moving away from promotional analysis that happens primarily after an event toward scenario modeling that can influence decisions before a promotion launches.

For retailers, the objective is not simply greater forecast accuracy. Better predictions can potentially support fewer stock-outs, more efficient inventory allocation and improved promotional economics.

Strategic Outlook

IKI's implementation demonstrates how AI can turn promotional planning into a more predictive discipline.

The immediate question is whether the reported 10-plus-point improvement in selected categories can scale across the retailer's wider assortment and different promotional scenarios. Grocery demand is highly variable, so performance during one category or period does not automatically establish a universal improvement.

Longer term, the more significant opportunity is integration. Promotional AI can become more valuable when its forecasts connect with replenishment, pricing, merchandising, customer analytics and broader retail planning systems.

That creates a model in which marketing decisions are informed by operational consequences before the promotion reaches the shopper.

Top Insights

• IKI reports more than a 10-point improvement in promotional forecast accuracy in selected categories after adopting SymphonyAI's AI-driven planning capabilities.
• Promotional AI shifts planning from hindsight to prediction, allowing retailers to model expected sales lift, margin impact and product cannibalization before campaigns launch.
• Marketing and supply-chain decisions are becoming connected, because promotional demand directly affects inventory requirements and product availability.
• IKI's 250-store network provides a substantial deployment environment, giving the retailer an opportunity to evaluate whether initial results can scale across additional categories.
• AI forecasting is becoming a broader retail infrastructure layer, with applications extending from promotional planning to replenishment, allocation and demand forecasting.

Get in touch with our MarTech Experts.