Building Smarter Cities Through AI-Driven Digitization: A Case Study

Zuzana Schwarzov

, Leonard Walletzk

, Mouzhi Ge

and Patrik Proch

azka

Faculty of Informatics, Masaryk University, Brno, Czech Republic

European Campus Rottal-Inn, Deggendorf Institute of Technology, Deggendorf, Germany

{433529, 133}@muni.cz, mouzhi.ge@th-deg.de, 418277@muni.cz

Keywords:

Digitization, Smart Cities, Artiﬁcial Intelligence, Municipal Services, Document Intelligence.

Abstract:

The concept of Smart (or Smarter) Cities is widely recognized in contemporary society. The integral relation-

ship between Smart Cities and digitization has been extensively researched, establishing it as a fundamental

condition for developing modern, effective, and sustainable services within the intricate environment of a

Smart City. This paper focuses on the implementation methods of digitization. Numerous cities are transition-

ing their agendas from traditional (analogue or paper-based) to digital platforms. However, the impact of such

digitization can vary signiﬁcantly. In many instances, cities develop a one-to-one digital replica of an analogue

or paper service, neglecting to explore the potential for improved service utilization or integration with other

services. This often overlooks the opportunity to leverage synergistic effects that could enhance value for all

stakeholders, including citizens, administration, and business entities. We aim to investigate the digitization

process at the Ofﬁcial Board of a municipality, using the results from a Hackathon organized in Brno, Czech

Republic, as an example of an innovative approach to such solutions. We intend to discuss the methods of

digitization provision and based on the case study suggest a best practice approach to avoid common mistakes

and issues arising from an incorrect approach to digitizing public services.

1 INTRODUCTION

The concept of smart cities has gained considerable

attention in recent years, driven by rapid urbanization

and the need for sustainable urban development. A

smart city uses technology and data analytics to im-

prove the quality of life for its citizens (Adje et al.,

2023), enhance urban services, and ensure efﬁcient

resource management. Digitization, the process of

converting information into digital formats, is a crit-

ical component of smart cities. It enables system

integration and real-time data analytics for decision-

making (Belli et al., 2023).

Digitization encompasses a broad spectrum of

technologies and applications, including Internet of

Things (IoT) devices, sensors, data analytics plat-

forms, and artiﬁcial intelligence (AI) (Dhiman and

Alghamdi, 2024). These technologies collectively

contribute to smarter infrastructure, improved public

services, and environmental sustainability. For exam-

ple, IoT devices and sensors continuously collect data

on urban parameters such as air quality, trafﬁc ﬂow,

and energy use (He et al., 2023). This data is pro-

cessed using analytics platforms to enable optimized

decision-making. AI further enhances smart cities

by providing predictive analytics and optimizing re-

source allocation (Lehti

o et al., 2023), such as pre-

dicting trafﬁc congestion patterns (Jiang et al., 2022)

and balancing energy supply and demand and the ef-

ﬁcient use of renewable energy sources (Zavorka and

Paar, 2022).

Traditional urban management approaches often

fall short in addressing the complexities of smart

cities. Therefore, studying how digitization con-

tributes to smarter cities is essential. Through a case

study, we can gain insight into best practices, identify

potential barriers, and propose strategies for possible

implementation. This paper explores digitization as a

critical building block in smarter cities through a case

study in the Czech Republic. Speciﬁcally, it examines

the role of digitized services in the ofﬁcial boards of

cities and municipalities and how digital technologies

and AI are integrated into urban planning and man-

agement. The paper discusses the challenges faced in

the digitization process and the opportunities for fu-

ture urban development.

The rest of the paper is organized as follows: Sec-

tion 2 reviews the relationship between smart cities

and digitization, especially the SMART principles in

the context of the Czech Republic. Section 3 ad-

172

Schwarzová, Z., Walletzký, L., Ge, M. and Procházka, P.

Building Smarter Cities Through AI-Driven Digitization: A Case Study.

DOI: 10.5220/0013478700003953

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2025), pages 172-179

ISBN: 978-989-758-751-1; ISSN: 2184-4968

dresses digitization priorities from both the city and

citizen perspectives. Section 4 conducts a case study

on ofﬁcial board digitization, including identiﬁed is-

sues, proposed solutions, and future plans. Finally,

Section 5 concludes the paper and outlines lessons

learned from the case study and future works.

2 SMART CITIES AND

DIGITIZATION

Smart cities rely heavily on information and commu-

nication technology (ICT). Effective use of any tech-

nology requires data, which can be derived from the

city’s behavior, stakeholder inputs, and city agendas.

The trend of smart city services is closely linked to

digitizing all city processes.

The challenges of digitization are well reﬂected

by the service researchers. Many of them intercon-

nect the value of digitization with urban develop-

ment (Bayat and Kawalek, 2023) and as the aspect of

the globalization of the public sector (L

apple, 2001).

There are two aspects of the digitization of city ser-

vices: the digitization of the public services (aimed at

citizens or other other service consumers) and the dig-

itization of inner services (processes) of the city. The

success of this process is critical for the acceptance

of digital services by the public, even if they could be

very innovative and potentially valuable (Rajagopalan

and Ravi, 2020).

It is important to note that digital service is just

one perspective of how we can look at the smart city,

or it is one context from many others because a smart

city is a typical multi-contextual environment (Wal-

letzk

y, Leonard et al., 2023). Therefore, digitization

can be an essential part of smart city services, reﬂect-

ing the needs of agents from other domains or con-

texts, where those contexts can play a major role in

the acceptance of the ﬁnal digital service (Badr et al.,

2021). The digitization of the smart city services is

a typical example of a service that directly inﬂuences

the willingness of stakeholders e.g. typically citizens,

but also tourists, business entities etc. to collaborate

with the municipality and adapt their behavior for bet-

ter value co-creation (Polese et al., 2018). Therefore,

the digitization of city services needs reﬂect a wider

context of the city environment and collaboration with

the whole landscape of academia, government, indus-

try, and public sector (Walletzk

y et al., 2022).

2.1 Application Context in Czechia

This paper explores a case study based on the con-

text of the Czech Republic. That is why we present

the Czech Republic’s latest ofﬁcial methodology for

smart city development, titled as Smart Cities Con-

cept - resilience through SMART solutions for mu-

nicipalities, cities, and regions. It was issued by the

Ministry of Regional Development of the Czech Re-

public, and approved by the Czech government on the

May 2021 (Ministry of Regional Development,

2021).

This methodology takes into account Czechia’s

uncommon settlement structure, which is formed by

more than 6000 municipalities. The concept also

considers that the municipalities are scaled differ-

ently, and addresses the need to faciliate innovative

solutions that would be valuable for all scales, al-

together with the public administration at all levels.

It also emphasizes citizen-centric approach and its

goal is to raise the citizen’s quality of life, quality

of available services and to create good living condi-

tions regardless of the size or location of the munici-

pality (Ministry of Regional Development, 2021). To

achieve this goal, the concept uses smart solutions,

which represent innovative approaches to problem-

solving in municipalities. A solution can be SMART

only if it respects the SMART principles, which

are a base for the whole concept. There are seven

SMART principles, which are translated from Czech

by (Schwarzov

a, 2023):

1. The Principle of Direction Change: this means

creating the conditions so that, where it is possible

and efﬁcient, services are delivered to people, and

work and business can be carried out from home

or a place close to home.

2. The Principle of Resilience: this is the resilience

of people and communities, the local economy,

the environment, and cohesion in the territory

based on digitization and innovative solutions.

3. The Principle of One Solution with Multiple

Effects: a solution is expected that will bring sev-

eral signiﬁcant effects (solving multiple needs at

once) with a holistic approach.

4. The Principle of “Short Distance”: everything

that can be provided locally must be provided lo-

cally or at the shortest possible distance (using

rule 3E - economy, efﬁciency, effectiveness).

5. The Principle of Cooperation and Financial

Sustainability for the Aim of achieving Effec-

tiveness of the Solution: it is about coopera-

tion with all partners in the territory, the usage

of multi-source ﬁnancing regarding its long-term

sustainability.

6. The Principle of Cohesion and Complementar-

ity, Horizontal and Vertical Interconnection:

Building Smarter Cities Through AI-Driven Digitization: A Case Study

173

the new solution leads to a leveling of opportuni-

ties, reduces tensions, solutions follow each other,

cooperation and interconnection at all levels and

all levels of public administration is a basic pre-

requisite for achieving resilience and cohesion.

7. The Principle of Evidence-Based Solutions

Based on Facts, Openness and Data Sharing,

Transparency, and Equal Opportunities: data

is generated that is understandable and accessi-

ble for innovative applications and the develop-

ment of people’s lives, communities, and busi-

nesses (Sharing is caring).

The structure of the concept is layered (Figure 1),

with the SMART solutions on the top, as an overlap-

ping roof above all the other parts. Below them is a

covering cross-sectional area, connecting the SMART

solutions together. It is titled the resilience, a direct

link to the name and goal of the concept. Under these

layers, there are three pillars of sustainable develop-

ment, representing the most important elements - peo-

ple and communities, local economy, environment for

living. These pillars focus on achieving resilience

through the citizen’s point of view. Each of these

pillars and the cross-sectional area is furthermore di-

vided into 4 components with different focuses. The

Smart Cities Concept recognizes digitization as one

of the integral parts of building smarter cities, includ-

ing it as one of the components of the Pillar B (Lo-

cal economy), which is related to competitiveness of

Czechia’s cities, municipalities and regions. Our pa-

per highlights the components of Pillar B: (1) Busi-

ness is a natural part of the life of a municipality, city,

and region, (2) Citizens and municipalities / cities / re-

gions as energy supplier partners, (3) Raw materials

and recycled materials in the circular economy, de-

velopment of the bioeconomy, (4) ICT infrastructure

- a basic prerequisite for the success of digitization

(Ministry of Regional Development, 2021).

Figure 1: Structure of the Smart Cities Concept. Based on

(Ministry of Regional Development, 2021), (Schwarzov

2023).

The fourth component is directed on the impor-

tance of having a sufﬁcient infrastructure for the in-

formation and communication technologies (ICT). It

sets the infrastructure as a prerequisite for the success

of digitization and aims to achieve three set goals:

(1) Sufﬁciently sized ICT infrastructure is available

throughout the whole territory of the Czech Republic.

(2) Cities, municipalities and regions have the nec-

essary infrastructure and ICT equipment for their ac-

tivities. (3) The ICT infrastructure enables the safe

development of digital services at the level of cities,

municipalities and regions.

These goals are supported by actual implementa-

tion of SMART solutions in the municipalities, cities

and regions (Ministry of Regional Development,

2021). The process of services digitization in

a municipality can encompass multiple domains

- healthcare services, social services, transport

services, etc. In this case study, we are focusing

on the digitization of government services, more

speciﬁcally, the local administration services in cities

or municipalities.

2.2 Government Services Digitization

Digitizing government services provides citizens with

quicker and easier access to information, enabling

them to be better informed about city processes (Neis

et al., 2023). A key focus is on creating an environ-

ment where citizens can participate in the co-creation

of public services in their municipality (Viale Pereira

et al., 2018).

Findings of (Twizeyimana and Andersson, 2019)

suggest that the usage of digitized services in form

of e-government can improve the public services, ad-

ministration and through that, it enhances the overall

well-being of the society and citizens. The authors of

(Neis et al., 2023) further build on that analysis, and

present possible advantages and challenges in digiti-

zation of government services. Firstly, for the advan-

tages, it can lead to the improvement of public ser-

vices and administrative efﬁciency, which can encom-

pass for example increased quality of the services,

boosted responsiveness to citizen’s needs, reduction

of administrations costs or more efﬁcient data storage.

Using automated electronic services that remove the

human link can also help with minimizing the pos-

sibility of corruption, more ethical behavior and in-

creasing the trust in administrations services.

In order to utilize digitized government services

to their full potential, it is also needed to be aware of

the possible challenges. The challenges do not need

to pertain solely to the technical process of digitiza-

tion. A signiﬁcant challenge stems from the digital

literacy of citizens. In digitizing the services in lo-

cal administration that are needed for every citizen,

there is a need to consider the citizens that can lack

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

174

the minimal digital literacy, for example the usage of

the internet. This can create a problem called digital

exclusion. The services also need to be planned ac-

cessibly, to bring equal value to citizens with various

disabilities. Furthermore, transparency of the service

processes also plays a role, and can lead to more ac-

tive participation of communities. As for the admin-

istrative digitization itself, it needs to adhere to the

legislative and policies, and stay up to date with the

current laws. Citizens’ data, their personal informa-

tion, must be protected by sufﬁcient security to build

the trust of citizens (Neis et al., 2023).

3 PRIORITIES OF

DIGITIZATION

3.1 Point of View from Cities

From the city’s perspective, there are two main view-

points: politicians and administration.

In the politicians’ perspective, they are ﬁrmly fo-

cused on everything that can be sufﬁcient for the cit-

izens. Not because they primarily reﬂect their posi-

tion, but because they want to be elected again. There-

fore, many politicians tend to shorten the digitization

to portals for citizens or services for the citizens. In

many cases, they do not understand the links and rela-

tionships between services and underestimate the ad-

ministrative and process support of the services for

citizens. Also, because they are trying to act as ex-

perts (but they are not), they often support the ser-

vice with a very limited or hardly achievable effect.

This can result in unsuccessful services that citizens

refuse to use. For example, in one Czech municipal-

ity, politicians decided to create their own portal for

citizens to make access to city services easier. The

registration process was so complicated that almost

all citizens refused to use the application.

Administrators, on the other hand, mostly employ

a rational approach and see digitization as a com-

plication, requiring changes to existing IT systems

and processes. There is also a typical problem - so-

called ”resort-ism”, the tendency to view problems

only from their speciﬁc department’s perspective and

not holistically. This is the source of many problems

and issues hidden from the view of citizens.

Even though it could be that digitization has many

obstacles in the cities, the truth is the opposite. The

previous two bullets present a typical ”negative” ap-

proach to digitization; every service designer (or ar-

chitect) should reﬂect it. There are also many exam-

ples of a very positive approach and synergy among

the politicians and administration that support digitiz-

ing and creating new, innovative applications. How-

ever, this process is susceptible to the support of both

sides - politicians and administration.

3.2 Point of View from Citizens

The citizens’ point of view is a little bit easier. They

want to get an effective tool that enables them to

get all information and any interaction with the city

quickly, effectively and with fewer resources. This is

also the problem - because they must accept the digi-

tal services of the city, they must understand the value

proposition, and they must have the capabilities to

participate in value co-creation (Caputo et al., 2017).

Many municipalities and their representatives do not

take into consideration that citizens or generally the

ﬁnal receivers of the services must also use their re-

sources to create the collaboration path for the value

co-creation (Polese et al., 2018) and only way how to

achieve this is to formulate proper value proposition -

understandable and acceptable for the citizens. Then,

the citizens will be willing to use their resources (in-

formation and knowledge) to participate in the value

co-creation. The other aspect of digitization is also

the social dimension, where many people can feel

less comfortable or have problems trusting digital ser-

vices. That also leads to ethical problems that can also

harm the successful digital service that is not thor-

oughly followed with the ethical consequences (Badr

et al., 2021).

4 CASE STUDY: OFFICIAL

BOARD DIGITIZATION

4.1 Legislative in Czechia

The ofﬁcial board in cities and municipalities can

be deﬁned as a publicly accessible surface on which

the administrative ofﬁce publishes documents, such

as legal regulations, municipality decisions and other

documents of administrative ofﬁces or courts. These

published documents are usually in paper form. In

the Czech Republic, the administrative ofﬁces are re-

quired by law to have an ofﬁcial board. The ofﬁ-

cial board needs to be: accessible nonstop, secured

against unauthorized handling of documents, avail-

able through a remote access, marked as an ofﬁcial

board and contain the contact information for the ad-

ministration ofﬁce responsible .

”Electronic” ofﬁcial board is not deﬁned in the

legislative. It is considered to be the fulﬁllment of

Building Smarter Cities Through AI-Driven Digitization: A Case Study

175

a legal obligation to publish the content of the ofﬁ-

cial boards in a way that allows remote access. In

most cases, the administrative ofﬁce creates web-

pages where they simultaneously publish the content

of the physical ofﬁcial board in an electronic form

(Ministry of the Interior, 2009).

4.2 Case Study

The study of this article focuses on the city of Brno,

which is situated in the southern part of Czechia. Its

ofﬁcial number of residents with permanent residence

is more than 400 000 people, however, during the

working week there are approximately 500 000 peo-

ple in the city (Data Brno, 2024a). The city itself

is divided into 29 self-governing municipal districts,

each with their own mayor, council, and coats of arms

(City of Brno, 2024b). The overarching and highest

government body for all city districts is the Brno City

Assembly (City of Brno, 2024a).

The ofﬁcial boards are an integral part of the com-

munication from the city towards the citizens. How-

ever, there is often a problem with their usability and

user friendliness in the remote electronic form. The

current management of the City of Brno wants to

improve on the electronic ofﬁcial boards, which led

them to connect the topic of their modernization with

this year’s city Hackathon, called Hack Brno. This

Hackathon was organized by the city of Brno with the

cooperation of the non-proﬁt organization Czechitas

and Brno.AI platform. The main challenge and as-

signment was to create an application, tool, or anal-

ysis which would be based on the open data of the

city and which would bring value to the city and its

residents. The participants formed teams and had 24

hours to work on their projects. The winning team,

in which one of the authors has participated, created a

proof-of-concept solution that has brought several im-

provements to electronic ofﬁcial boards (Data Brno,

2024b).

4.2.1 Identiﬁed Problems

At ﬁrst, the team has analyzed the current state of

the ofﬁcial board of Brno. Brno City has its ofﬁcial

board (City of Brno, 2024c) and furthermore, each of

the 29 municipal districts has its own ofﬁcial board

that can differ in style (Municipal district Brno - Jih,

2024; Municipal district Brno -

Zabov

resky, 2024).

That means that there are 30 ofﬁcial boards in oper-

ation simultaneously. The current electronic ofﬁcial

boards meet the minimum requirements mandated by

legislation, which were listed in the previous section.

They publish the ofﬁcial documents in a PDF format,

accompanied with the relevant metadata structured as

Region, Category, Title, Reference number, Origina-

tor and the Publication period of the document. The

users can ﬁlter through these documents based on

chosen metadata. The electronic ofﬁcial boards also

provide a search option as a result of all the document

contents.

4.2.2 Proposed Solution

Based on the assignment of the hackathon, the pro-

posed solution utilizes the available artiﬁcial intelli-

gence technologies to provide intelligent data extrac-

tion and analysis. It aims to counter the overﬂow of

ofﬁcial boards in Brno by congregating the data in one

board that will be searchable.

In the hackathon project, signiﬁcant improve-

ments were made to the user interface to enhance us-

ability and ensure a consistent experience across all

ofﬁcial boards. While many of these enhancements

focused on the technical aspects of the project, the in-

terface redesign aimed to create a more user-friendly

and uniﬁed experience.

A main task and challenge of the hackathon was

the effective utilization of artiﬁcial intelligence. This

was accomplished through the implementation of

Azure AI Document Intelligence (DI) and Azure AI

Search. Given the tight 24-hour timeframe of the

hackathon, we selected Azure services to speed up the

development process and allow us to focus more on

providing the solution rather than tackling the com-

plexities of model ﬁne-tuning, data indexing, search

engine tuning, and infrastructure setup. Utilizing

Azure’s prebuilt AI services enabled rapid develop-

ment and deployment, which was crucial under the

hackathon’s time constraints.

The DI service was used to convert PDF docu-

ments into a structured JSON format, a crucial step

given the variability in document templates across

different districts. By deﬁning a standardized JSON

structure, the project was able to eliminate these dis-

crepancies and unify the data. Creating custom mod-

els for DI was necessary to achieve this data uniﬁ-

cation. These models were trained to extract the re-

quired information from each document and populate

the corresponding ﬁelds in the JSON objects. The re-

sulting JSON objects were then stored in a database.

Training custom models, instead of using a pre-

trained one, for our use case was necessary, but it in-

troduced additional complexity to the project. The DI

requires at least ﬁve training documents to be labeled

by hand for each document type, and there are numer-

ous document types, such as city inquiries, lost and

found notices, announcements, public decrees, job of-

fers, etc. Each document type has a speciﬁc structure,

and each municipality may have different templates

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

176

for the same category. This variability makes training

custom models to achieve acceptable accuracy and

conﬁdence scores a signiﬁcant challenge.

Figure 2: Example of Training Data for Document Intelli-

gence - Lost and Founds.

In Figure 2, various colors are used to represent

different categories of information extracted from the

document. For example, red indicates a contact per-

son, orange denotes a phone number, green signiﬁes

an email contact, blue represents the date when the

items were found, and purple highlights the items

listed in the document. The number and types of la-

bels in the document may vary depending on the cat-

egory. Conﬁdence scores are provided after building

a custom model and include (Microsoft, 2024):

• Document Type Conﬁdence Score: The docu-

ment type conﬁdence is an indicator of closely the

analyzed document resembles documents in the

training dataset. When the document type con-

ﬁdence is low, it is indicative of template or struc-

tural variations in the analyzed document.

• Field Level Conﬁdence: Each labeled ﬁeld ex-

tracted has an associated conﬁdence score. This

score reﬂects the conﬁdence of the model on the

position of the value extracted.

• Word Conﬁdence Score: Each word extracted

within the document has an associated conﬁdence

score. The score represents the conﬁdence of the

transcription.

• Selection Mark Conﬁdence Score: Each selec-

tion mark has a conﬁdence score representing the

conﬁdence of the selection mark and selection

state detection.

Using a custom model for each category seemed

to be the optimal approach. This strategy led to

achieving conﬁdence scores of over 85% with fewer

than 10 documents as training data. While this score

was sufﬁcient for the proof of concept, for produc-

tion use across all 30 municipalities, it is imperative

to train on a larger number of documents per category

to accommodate the discrepancies in documents from

different municipalities.

With the custom models created, built, and de-

ployed, the API served as the primary means of com-

munication between the Document Intelligence (DI)

service and the application. The response from the DI

service is a large JSON object that includes content

for every label detected in the document, along with a

conﬁdence rating. This JSON is then transformed by

the mapper into the resulting JSON structure.

The source structure for this JSON is deﬁned in a

Java class corresponding to each category. As an ex-

ample, the class for the job offer category is depicted

in Figure 3. The structure of the class varies for each

category. These classes were created, and the ﬁnal-

ized content for each category should be reviewed and

discussed with stakeholders.

@Data

public class JobPosting {

private String job_position;

private String job_description;

private String job_contact_person_name;

private String job_contact_person_phone;

private String job_salary_condition;

private String job_start_date;

private List<String> requirements;

private List<String> job_benefits;

private List<String> application_requirements;

private String job_application_description;

}

Figure 3: JobPosting Class Deﬁnition.

Azure AI Search service was implemented over

ﬁnalized JSON data, enabling intelligent search ca-

pabilities. This service allowed users to perform

searches across documents using natural language

queries. For instance, a query such as ”I lost my cell

phone” would return a list of records from the ofﬁcial

boards categorized under lost and found, speciﬁcally

identifying documents that mentioned a phone in the

attachments. The application can be deployed with-

out requiring any actions from the municipality and

can be used simultaneously with the current solution.

Data for the project are scraped from the ofﬁcial API

of the ofﬁcial boards, ensuring seamless integration

and up-to-date information without disrupting exist-

ing workﬂows.

4.2.3 Further Plans

The team’s results from the Hackathon were sufﬁ-

cient as a proof of concept, and the evaluators’ feed-

back highlights the potential for a nationwide deploy-

ment of such upgraded electronic ofﬁcial boards for

the future. However, several challenges need to be

Building Smarter Cities Through AI-Driven Digitization: A Case Study

177

addressed before moving forward.

To address these challenges, we decided to cre-

ate the data extraction process with custom logic in-

stead of relying on multiple models within Document

Intelligence (DI) tools. This approach mitigates the

high costs associated with DI tools while providing a

more ﬂexible and efﬁcient solution tailored to our use

case. Since using commercial AI tools can be costly,

identifying open-source alternatives can enhance efﬁ-

ciency in the long-term operation of the service. An-

other signiﬁcant challenge is the variability in docu-

ment structures across different categories and munic-

ipalities. This can be done by standardizing the docu-

ment structures, which would be advantageous for all

stakeholders. Also, it would enable citizens to have a

consistent understanding of what to expect in ofﬁcial

documents.

We have moved beyond the commercial tools dur-

ing the Hackathon proof-of-concept phase to ensure

greater ﬂexibility and alignment with the project’s

long-term goals. Further, we are developing the sys-

tem entirely on open-source technologies, which al-

lows for deeper customization and integration explic-

itly adapted to our use case.

Optimizing the data extraction process from PDF

documents is one key objective, which is to address

complex elements such as tables and images often

present in municipal board notices. The challenge lies

in transforming tabular data into textual formats that

is suitable for embeddings. The process requires con-

verting the structured data from tables into coherent

text representations while preserving semantic rela-

tionships. These text representations will then be en-

coded into embeddings, ensuring that the contextual

and relational information from the original tables re-

mains intact and meaningful for downstream tasks.

Besides table parsing, the focus will be on detect-

ing and correcting errors in the extracted text. The

validity and accuracy of text involve implementing a

classiﬁcation system to identify and ﬁlter out invalid

extractions, such as incomplete words, invalid charac-

ters, or non-text elements. These improvements will

contribute to a more reliable and accurate data extrac-

tion process. The revised and corrected text will be

used to generate sentence-level embeddings for se-

mantic search. These embeddings will be linked to

the original text segments to keep traceability and ef-

fective information retrieval within the system. Fur-

thermore, applying representative search queries’ sce-

narios can be used to test the system’s semantic search

capabilities. These queries evaluate the system’s abil-

ity to retrieve relevant results from the embedding

database, which provides a reliable benchmark for se-

mantic search performance.

5 DISCUSSION

In the last section, we showed the development of an

application that could be a typical example of digiti-

zation. However, we have also revisited many aspects

that may block its successful deployment to the ser-

vice structure of the municipalities. Also, the usage

of AI and machine learning presents a game changer

in different aspects of human life. We may always

consider the issue related to the typical resistance to

changes. Overcoming resistance to change within or-

ganizations in the development process is essential to

ensure effective implementation.

The case study we presented was highly success-

ful. It wins the ﬁrst prize at the hackathon and

receives positive feedback from public institutions.

There were even offers to deploy the solution in mu-

nicipalities. However, some challenges remain, such

as ﬁnishing the student-developed application, decid-

ing who will maintain and improve it, and consider-

ing if and how the university should be involved in the

process.

Digitization of municipalities and states brings

many possibilities but also questions and obstacles.

We stand before a new situation where universities

should play a more active role, not only supporting

research transfer but also helping students ﬁnd inno-

vative ways to participate in research as partners with

municipalities and society.

The development of the digitized application for

municipal ofﬁcial boards highlights how municipal-

ities handle and share public data. The project has

seen initial success but faces challenges related to de-

ployment and long-term sustainability. Differences

in document structures across municipalities may re-

quire multiple custom models for document process-

ing. The cost of commercial AI tools makes it es-

sential to consider open-source alternatives. Since the

success of such a digitization project depends on the

involvement of city stakeholders, their active engage-

ment is important for continuous improvement and ef-

fectiveness of the application.

6 CONCLUSION

In this paper, we have studied the development and

implementation of a digitized application for munic-

ipal ofﬁcial boards. This case study demonstrates its

initial success for public service efﬁciency and acces-

sibility. The hackathon project has proven the feasi-

bility of this application. Some lessons learned can

be derived as scalable model training, cost-effective

AI solutions, and stakeholder engagement. Further,

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

178

it can be seen that a close partnership between uni-

versities, municipalities, and technology providers is

essential for innovations.

As future work, we plan to reﬁne the data extrac-

tion process with custom logic rather than relying on

multiple models within Document Intelligence tools.

This decision aligns with our goal to mitigate costs

and ensure greater ﬂexibility.

REFERENCES

Adje, K. D. C., Letaifa, A. B., Haddad, M., and Habachi,

O. (2023). Smart city based on open data: A survey.

IEEE Access, 11:56726–56748.

Badr, N., Walletzk

y, L., Dr

agoicea, M., Carrubbo, L., and

Toli, A. M. (2021). Modelling for ethical concerns

for traceability in time of pandemic “do no harm” or

“better safe than sorry!”. In Proceedings of the 54th

Hawaii International Conference on System Sciences,

pages 1779–1788, Honolulu, Hawaii. HICSS.

Bayat, A. and Kawalek, P. (2023). Digitization and urban

governance: The city as a reﬂection of its data infras-

tructure. International Review of Administrative Sci-

ences, 89(1):21–38.

Belli, L., Davoli, L., and Ferrari, G. (2023). Smart city as

an urban intelligent digital system: The case of parma.

Computer, 56(7):106–109.

Caputo, F., Ge, M., and Walletzk

y, L. (2017). Modelling

the service value chain for smart city. In Proceedings

of the 25th Interdisciplinary Information Management

Talks, pages 375–382, Pod

ebrady, Czech Republic.

Trauner Verlag.

Dhiman, G. and Alghamdi, N. S. (2024). Smose: Arti-

ﬁcial intelligence-based smart city framework using

multi-objective and iot approach for consumer elec-

tronics application. IEEE Trans. Consumer Electron.,

70(1):3848–3855.

He, X., Ai, Q., Wang, J., Tao, F., Pan, B., Qiu, R. C., and

Yang, B. (2023). Situation awareness of energy inter-

net of things in smart city based on digital twin: From

digitization to informatization. IEEE Internet Things

J., 10:7439–7458.

Jiang, Y., Han, L., and Gao, Y. (2022). Artiﬁcial

intelligence-enabled smart city construction. J. Super-

comput., 78(18):19501–19521.

apple, D. (2001). City and region in an age of globalisation

and digitization. German Journal of Urban Studies,

40(2):13–34.

Lehti

o, A., Hartikainen, M., Ala-Luopa, S., Olsson, T., and

anen, K. (2023). Understanding citizen percep-

tions of AI in the smart city. AI Soc., 38(3):1123–

1134.

City of Brno (2024a). City authorities. https://en.brno.cz/

w/brno-city-authorities.

City of Brno (2024b). Municipal disctricts. https://www.

brno.cz/mestske-casti-prehled.

City of Brno (2024c). Ofﬁcial board - news. https://edeska.

brno.cz/eDeska/eDeskaAktualni.jsp.

Data Brno (2024a). Brno in numbers - citi-

zens and housing. https://data.brno.cz/pages/

bvc-obyvatelstvo-a-bydleni.

Data Brno (2024b). Hackathon 2024. https://data.brno.cz/

pages/hackathon2024.

Ministry of Regional Development (2021). Smart cities

concept: resilience through smart solutions for munic-

ipalities, cities and regions. https://www.budtesmart.

cz/koncepce-smart-cities.

Ministry of the Interior (2009). Methodological rec-

ommendation for the activities of territorial self-

governing units. https://www.mvcr.cz/odk2/soubor/

modra-c-4-web-pdf.aspx.

Municipal district Brno - Jih (2024). Ofﬁcial

board: Municipal district brno - jih. https:

//www.brno-jih.cz/vismo/zobraz\

dok.asp?p1=

55\&id\ org=80907\&ud=1.

Municipal district Brno -

Zabov

resky (2024). Ofﬁcial

board: Municipal district brno -

Zabov

resky. https:

//www.zabovresky.cz/uredni-deska/.

Microsoft (2024). Interpret and improve model accuracy

and analysis conﬁdence scores.

Neis, M. I., Benitez, G. B., and Frank, A. G. (2023).

Smart cities and the digitization of government ser-

vices: Beneﬁts and challenges. In Proceedings of

the 11th International Conference on Production Re-

search – Americas, pages 16–24.

Polese, F., Barile, S., Caputo, F., Carrubbo, L., and Walet-

zky, L. (2018). Determinants for value cocreation and

collaborative paths in complex service systems: A fo-

cus on (smart) cities. Service Science, 10(4):397–407.

Rajagopalan, S. and Ravi, S. (2020). A critical appraisal of

digitalization in city administration. In AMCIS.

Schwarzov

a, Z. (2023). Smart solutions, concept structure:

Seslab working material.

Twizeyimana, J. D. and Andersson, A. (2019). The public

value of e-government – a literature review. Govern-

ment Information Quarterly, 36(2):167–178.

Viale Pereira, G., Eibl, G., and Parycek, P. (2018). The role

of digital technologies in promoting smart city gov-

ernance. In Companion Proceedings of the The Web

Conference 2018, WWW ’18, page 911–914.

Walletzk

y, L., Nenckov

a, L., and Schwarzov

a, Z. (2022).

Management of smart city services from the perspec-

tive of quadruple helix. In ICERI2022 Proceedings,

15th annual International Conference of Education,

Research and Innovation, pages 3158–3163. IATED.

Walletzk

y, Leonard, Carrubbo, Luca, Ge, Mouzhi,

Schwarzov

a, Zuzana, and Bayarsaikhan, Odonchimeg

(2023). Multi-contextual smart city model for service

interconnections. ITM Web Conf., 51:01001.

Zavorka, R. and Paar, M. (2022). A focus on electromobil-

ity within smart city solutions - charging stations, re-

newable energy, and air quality monitoring. Sensors,

22(20):7841.

Building Smarter Cities Through AI-Driven Digitization: A Case Study

179