Retail customers need better demand forecasts to ensure that there is no build-up, and eventually waste of excess stock and to avoid stock-out situations.

Using relevant datasets such as historical sales, product, location and google trends data we built a machine-learning model that predicts sales for multiple products across different retailers for a predefined future time period.

This improvement in the forecasting process increases revenues by preventing stock-out situations, decreases costs by mitigating the build-up of excess stock and removing the need for discounting, and improves margins. We also find a decrease in manual work effort by automating the forecasting process. Overall, significant uplift in profitability and sustainability.

HR departments are constantly looking to reduce unconscious bias in hiring decisions but are not paying enough attention to reducing unwanted noise among hiring managers. In an ideal scenario all hiring managers would choose the same candidate for the role given a set of parameters, however this is not always the case.

We conducted a data-driven noise audit to identify sources of noise in the hiring process. We supplemented this with a machine learning algorithm comparing the qualifications of each candidate with those of current employees on similar roles that have received good performance reviews. Information about the candidate’s background such as the gender and ethnicity are excluded to avoid biased decisions. This generates a ‘suitability score’ for each candidate, helping managers filter through applicants and reducing bias in the decision-making process.

The benefits of this tool are three-fold: Firstly, it creates transparency about noise in the hiring process. Secondly it provides necessary support to (and not replacement of) hiring managers to speed up the hiring process and limits noise and unconscious bias in recruiting. Finally,implementing a data-driven hiring approach allows for better tracking of recruitment statistics such as whether diversity quotas are being met.

A lot of the companies we work with are interested in comparing their sustainability efforts to their competitors to understand where they rank and how they can improve.

We scraped relevant data from company websites, social media and financial information websites and developed an NLP algorithm that extracts sustainability keywords and analyses them to produce a brand sustainability score. This score is broken down into own company score (efforts, materials used), customer perception score (using social media comments) and ESG score.

Clearly identified opportunities to improve a company’s ranking and consequently support them to unlock more sustainable behaviours.

Financial services companies with online application processes can struggle with their conversion rates due to lack of personalised targeting.

We built AI tools that offer customers personalised customer journeys to maximise ROI. This is achieved by constructing an AI-powered routing system to ‘nudge’ customers in the right direction leading to greater conversion likelihoods.

Increased conversion by adapting the customer journey according to their demand and increased customer satisfaction​ by offering customer support when needed​, led to +20% earnings per year and +20% new business volume per year. Overall, significant uplift in profitability.

Customers often have little visibility on the impact & effectiveness of media campaigns. Media agencies that our customers work with often allocate budgets without end-to-end ROI transparency

We investigate daily media spend per channel, combined with daily number of clicks and conversions in the sales funnel to build a machine learning model from which identifies the optimal spend allocation per channel. We refine this over time with continuous A/B testing of spend shifts.

Increased media impact​through optimal budget allocation​, decreased customer acquisition cost​through optimal spend allocation​ and over 15%​ Increase in ROAS (Return on Advertising Spend)​. In some channels, up to 200% increase in conversion.Overall, significant uplift in profitability.

Companies are facing intense competitive pressure and growing costs whichlead to a reduction in margins; this reduces overall profitability

Using historical data on price changes, competitor pricing and specific customer characteristics, we calculate the price sensitivity per customer and use that to build a Decision Optimization Model. This model calculates the optimal price per customer segment to achieve margin and volume targets. A fully interactive dashboard allows our customer to examine the impact of varying certain parameters.

Increased profit margin​ through a smarter allocation of high prices and increased volume​ by calculation of prices based on the willingness to pay. Margin improvement by up to 30% with up to 40% increase in new business volume per year. Overall, significant uplift in profitability.

The allocation of salespeople and product consultants across retail stores is often inefficient and can lead to losses

We calculated the marginal utility of an additional day of salesperson presence per store and constructed a ranking algorithm to find the optimal store, date and person combination with the highest business value

Increased profits by identifying highest uplift potential​, greater transparency by analyzing factors that affect salesperson impact on sales​ and a reduction in costs by eliminating inefficient salesperson activities​, a reduction in emissions by limiting salesperson trips to the most impactful ones​. Revenue increases of up to 300%per year. Overall, significant uplift in profitability and sustainability.