AI audit at Slevomat and AI multi-agents for deal managers

Slowed-down processes due to manual creation of offers, data, and and texts. High administrative burden for deals managers and copywriters, routine work. Uneven quality of offers and occasional data errors.

Conducting an AI audit that meticulously mapped the company’s processes and identified over 50 specific opportunities for automation and efficiency improvement using AI for gradual deployment. Implementation of AI tools custom-tailored to the company’s needs, prioritizing the highest benefits.

Deal managers, copywriters

Continuous cooperation for over 1 year, implementation and expansion of the scope to other departments is still ongoing.

Both AI tools function as independent multi-agent systems connected to Slevomat’s systems via API. They also use third-party APIs, e.g., Google Places. The first automatically searches for establishment data using LLM and web scraping. The second generates deal text proposals according to specified data and templates. The tools are safe, easily integrable, and support work efficiency.

Head of Business Unit

Complete implementation

Supplier

Deal managers; Copywriters

The cost of generative artificial intelligence tokens

1. Multi-agent system for finding establishment details: A process that used to take tens of minutes is now a matter of a few clicks. Minimal risk of inaccurate information and customer complaints. The speed, accuracy, and reliability of the entire system led to an accelerated offer publication process and a better experience for both deal managers and Slevomat partners. 2. Multi-agent system for deal text creation: A task that used to take tens of minutes now becomes a task of a few minutes, without compromising the output quality. Acceleration of text creation by over 70%, reduced team workload, and more uniform output quality across the entire offer range. Sales representatives can focus on clients, not on writing. Copywriters no longer have to turn over every offer four times with supplementary questions for deal managers.

AI process audit Number of analyzed departments: 12 Number of interviews conducted: 30+ (across teams, not only with managers) Audit duration: 8 weeks Identified AI opportunities: 38 Prioritized proposals: 15 Tools recommended for the first pilot phase: 2 selected for deployment in the next phase of cooperation We are still evaluating the specific developed solution

Thank you for the great collaboration! The entire process from selection and onboarding, through planning and execution of all meetings, to the final outputs went perfectly. I can wholeheartedly recommend AIExcellence.cz. CCO | Slevomat Group Matěj Dvorský

cooperation

1. Multi-agent system for finding establishment details This tool automates one of the most time-consuming steps of the entire process: filling in information about establishments in individual offers. Deals managers previously spent tens of minutes manually searching for information from partner websites – verifying opening hours, service descriptions, or location details. This work was not only time-consuming but also prone to errors, as each supplier presents data differently. An example might be a sold voucher for a restaurant visit; the customer wants to use it – they do not make a reservation, they only look at the voucher information on Slevomat, where the opening hours information is inaccurate. The customer leaves with an unpleasant experience; a similar situation can occur with the address, secondary experiences/services, or prices. Therefore, this AI tool automatically fills in all key establishment data into Slevomat offers – for example, opening hours, address, contact, service descriptions, or other details. This now saves deals managers a lot of time, and the results are more accurate. 2. Multi-agent system for deal text creation The second solution is a multi-agent system that automatically suggests texts for individual offers – including descriptions, advantages, and key benefits. Previously, this work had to be done by the deals managers themselves, as they were closest to the details of each offer, or by copywriters. Our assistant utilizes structured data from the offer and generates a complete text proposal, including key marketing messages, according to the defined Slevomat style. For correct factual information, it searches the partner’s website and looks for important data there, which it processes into the form of a value proposition. The deals manager then only checks the content and possibly adds specific details. This means that the copywriter team can handle more text offers, or focus on more important and potentially more valuable offers.

Do not hesitate to implement AI in your company. It may soon happen that a competing company offers the same product at a price below your production costs thanks to AI automation of business processes.