- Webinar: Bridging the Data to Decision Divide Lanny Roytburg ­ Chief
  Commercial Officer - Cloverpop Oleg Royz ­ VP, Retail & CPG ­ DataArt
  Jan Mehmet ­ Exec Advisor - End July 14, 2025 www.cloverpop.com

Welcome!

Lanny Roytburg Co-Founder, Chief Commercial Officer · 15+ years of
growth strategy, foresight analytics and I&A transformation experience;
focused on CPG, Retail, Pharma and manufacturing · Co-founder Cloverpop

Oleg Royz VP, Retail & CPG, Data + AI · 25+ years driving tech enabled
growth through Data, Digital, and AI transformation. · Deep Retail, CPG,
and Distribution value chains industry expertise

Jan Mehmet Executive Advisor, Interim CCO · 25+ years within the Retail
industry leading Digital across global markets · Deep Retail expertise
across core functions including Sales, Marketing, Operations, Customer
service and Technology

Confidential & Proprietary

2

Amid a Dynamic, Complex and Fast Paced Business Environment, F500
Organizations Make \~10MM Decisions / Year

F500 Companies make \~10MM+ decisions a year... STRATEGIC DECISIONS with
long-term implications Companies make \~10M+ decisions per year

...While having to navigate greater complexity than ever before... In a
VUCA world...

Macro / Geopolitical Uncertainty

Supply Chain Disruptions

Less Predictability

Free Flow of Information

TACTICAL DECISIONS impacting day-to-day operations

Fast Evolving Consumer Needs

Turnover / Knowledge Drain

Confidential & Proprietary

33

Today's business environment is changing at a faster pace than ever
before, forcing organizations to re-assess decision-making practices

Speed of innovation, data and workforce dynamics are changing more
rapidly than even before...

... Moreover, The pace of AI adoption outstrips previous technological
revolutions, forcing faster adaptation

Agility ­ Combination of fast-evolving customer needs, regulations and
decreasing barriers to entry require organizations to accelerate GTM
decisions X-Functional Connectivity ­ Rigid, silo'd organizational
structures hinder speed and quality of decisions, lacking visibility
into xfunctional trade-offs

GenAI

Workforce dynamics ­ Increase in turnover, knowledge drain within
organizations pressure institutional knowledge and learning

Decisions must evolve to be faster, data-driven, and collaborative,
fostering innovation and adaptability in a dynamic business landscape

Confidential & Proprietary

4

The launch of ChatGPT in \`22 had set off a wave of excitement and
investment in GenAI capabilities

AI has taken center-stage...

CxO's are committing to efforts

Deployment is underway

Confidential & Proprietary

5

...And the Expectations have been massive

\[GenAI\] can automate up to 29.5% of work hours in the U.S. economy by
2030¹ Source: 1. McKinsey 2. MIT SLoan

Generative AI is projected to raise global GDP by approximately 7%¹

GenAI can improve the productivity of highly skilled workers by up to
40% compared to those not using AI²

Confidential & Proprietary

6

AI / Generative AI will have a transformative impact on future ways of
working, shifting where employees spend their time

AI will automate many day-to-day activities...

1 RETRIEVE Data

2 ANALYZE Data

3 ANSWER Questions

...shifting focus on higher order activities

4 SYNTHESIZE Insights

5 DEVELOP Recommendations

6 EXECUTE & Capture Value

Today TIME Spent

hsfn

As AI automates tasks and unlock access to intelligence...

...focus will shift towards creative, innovation and execution & value
capture

Tomorrow VALUE Delivered

Confidential & Proprietary

7

However, the initial use-case based approach had risk

DISPARATE GEN-AI USE CASES continue to emerge

Virtual Assistants

Knowledge Retrieval

Customer Chatbots Marketing Content Generation

BI Enhancements HR support

But there is a risk · Insufficient value-creation / ROI · Lack of
strategic focus & Interconnectivity · Unwieldly Technical debt · AI
fatigue / Wariness · Talent gaps · Security vulnerability We've been
here below

Confidential & Proprietary

8

Initial wave of deployments delivered mixed results

Just 1/4 of enterprises reporting significant value from their AI
initiatives¹ 25%

Share of companies pausing their AI initiatives / POC's² 42.0% 17.0%

Employee Levels of Burnout³

45.0%

38.0%

35.0%

Source: BCG AI Radar Survey

2024

2025

Frequent AI Users

Infrequent

Never

Source: S&P Global market intel.

Source: Quantum Workplace

Confidential & Proprietary

9

Companies are rethinking their AI strategies ­ "Time to get practical"

Confidential & Proprietary

10

Key lessons to take forward...

What

Failure Point 1: Hype over Value AI was launched in scattered pilots
without clear business priorities

Failure Point 2: Models Without Mindset Shift Sophisticated models
deployed without preparing people or adjusting process to use them

So-What

Most efforts fizzled, producing great demo's but little enterprise value

Adoption stalled as workflows stayed the same and teams lacked trust,
training and clarity to engage

NowWhat

AI must start from defined business outcome, value and strategic
alignment

Real impact will come from designing AI around people-- supporting how
they think, collaborate, decide and act

Failure Point 3: Information without Insight AI flooded teams with
content, and analysis--without clarity on what matters or what to do
next The signal got lost in the noise, leaving teams more informed but
no more aligned or effective AI must evolve from generating information
to surfacing insight-- focusing attention, framing options, and moving
work forward with purpose

AI creates real value when it's 1 anchored to business goals, 2 designed
for how people work, and 3 built to turn information into action.

Confidential & Proprietary

11

Aligning AI around the foundational building block of
business...Decisions

Decisions are where Data, People and Process come together to drive
action Data & Analytics

Human Experience /Expertise

Business Context / Process

Financial Performance = Decision Quality 95% Correlation between
decision practices and financials performance Source: Bain Global
Retooling Survey 2020 (n=953), over five-year period Confidential &
Proprietary

Organizations make \>10MM+ decisions each year - the most ubiquitous
activity STRATEGIC DECISIONS with long-term implications TACTICAL
DECISIONS impacting day-to-day operations 12

A Decision-Back Approach is Critical for Realizing Full Value of
Analytical Capabilities

A Decision-centric vision aligns D&A activity with stakeholder needs and
business objectives with a direct impact on quality of decision outcomes

Value Decision R eco mme ndat io n Insights DATA From starting with data

To starting with business value VALUE Decision R eco mme ndat io n
Insights Data

Without a decision-centric vision, CDAOs risk veering away from key
stakeholder needs and the imperative of driving better decision making,
not just better data. David Pidsley Analytics & AI Research Analyst,
Gartner

Confidential & Proprietary

13

What's in a "decision"

Business Goal In crea sed Revenue

Decision(s) To Be Made Should we expand distribution channels Which
whitespace should we enter? How do we optimize Media Spend ROI Marketing
Director

1.  Questions to ask

2.  Insights/Data/Analytics

Where is media currently underperforming against our business
objectives?

Perf. Mkt Lea d

Media performance dashboards

Which channels, audiences, or regions are delivering the strongest
return?

Perf. Mkt Lea d

Media performance dashboards

What trade-offs exist if we reallocate spend--brand impact, reach, or
retailer commitments?

S hop per Mkt

Retail media Marketing Mix

What is the forecasted impact?

Media Ana lytics

Forecast models (MMM, incrementality testing, A/B test results)

how do we operationalize it across our platforms and partners?

C ampaig n Exec.

4.  Reco & Decision Made

DSPs (The Trade Desk, DV360, Amazon Ads) Agency partners

Confidential & Proprietary

3.  Synthesis / Discussion 14

What's in a "decision"

1 Connect to business goals Business Goal

Decision(s) To Be Made

1.  Questions to ask

2 Aligned to how people work 2. Insights/Data/Analytics

3 Built to turn information into action 3. Synthesis / Discussion

In crea sed Revenue

Should we expand distribution channels Which whitespace should we enter?
How do we optimize Media Spend ROI

Where is media currently underperforming against our business
objectives? Which channels, audiences, or regions are delivering the
strongest return? What trade-offs exist if we reallocate spend--brand
impact, reach, or retailer commitments?

Perf. Mkt Lea d Perf. Mkt Lea d S hop per Mkt

Media performance dashboards Media performance dashboards Retail media
Marketing Mix

Marketing Director

What is the forecasted impact?

Media Ana lytics

Forecast models (MMM, incrementality testing, A/B test results)

how do we operationalize it across our platforms and partners?

C ampaig n Exec.

4.  Reco & Decision Made

DSPs (The Trade Desk, DV360, Amazon Ads) Agency partners

Confidential & Proprietary

15

End to End Decision Process

1

Decision Trigger

2 Decision Model & Orchestration

3

Decision Making

4

Decision Execution

5

Decision Learning

Ex: Should we change price?

Situation has changed / Event has occurred

Decision to-bemade identified

Consumer Respon se P&L Imp act Impact to brand Business Logic Defined

Elasticity

Fin. Analysis Brand Equity Study

Decision made

Data Collected, Synthesized, Recommendation accepted/rejected

Decision made, outcomes realized

Track, measure, leearn

Confidential & Proprietary

16

How AI is deployed

1 Decision Trigger

2

Decision Model & Orchestration

Ex: Should we change price?

Consumer Respon se P&L Impact

Impact to brand

Situation has changed / Event has occurred

Decision to-be-made identified

Business Logic Defined

3

Decision Making

Elasticity Fin. Analysis

Decision made

Brand Equity Study

Data Collected, Synthesized, Recommendation accepted/rejects

4

Dec is ion Execution

5

Decision Learning and Optimization

Decision made, outcomes realized

Decision Learning

Continuously monitor internal and external data for critical
signals--such as demand shifts, supply disruptions, or competitor
activity-- that prompt decision-making Tech stack: ML models for anomaly
detection, event- driven architecture, monitoring APIs

Support users in framing decisions through guided question generation
and structuring decision logic; routing drivers/questions to appropriate
stakeholders

Pull together structured and unstructured data sources to synthesize
inputs (Synthesis-AI) and generate insights and recommendations

Translate decisions into operational steps by creating or triggering
workflows across enterprise systems (e.g., pricing updates, media budget
shifts).

Tech stack: LLMs + structured prompt templates, retrievalaugmented
generation (RAG), context-aware UIs.

Tech stack: Multimodal data pipelines, predictive ML models, Neural
network, graph analytics, time- series forecast, simulation and
optimization

Tech stack: API orchestration layers, GenAI + agentic workflow logic

Source: Cloverpop

Confidential & Proprietary

Continuously learns from the outcomes of decisions, adjusting models and
improving future decisions based on feedback loops Tech stack:
Reinforcement learning, supervised ML retraining pipelines 17

Success Stories

Pharma

CPG

Consumer Health

Decision to be made Impact Results

What actions should we take to mitigate issues with the supply of \[raw
material\] to \[site\]J?an - Walmart Use Case? Life-saving drug with 130
unique inputs experienced raw material delivery delays, slowing
production & causing shortages. Critical to patient lives to make fast
decisions for alternate suppliers. +60% Faster time to decision \$9M
Revenue recovered

What action(s) can we take to improve brand performance? CPG company
wanted to drive improvements in brand performance via holistic Brand
Health analysis. Current output was expensive and not tracked against
actionable decisions. \~2X Faster delivery \~30% Cheaper vs traditional
suppliers

Where should we invest our media dollars to drive highest impact? Global
Consumer Healthcare brand spent hundreds of millions on media investment
­allocation process was lengthy, political, complex. 15+ data sources,
months of analysis to develop recommendations. 80% Acceleration in time
to insight & recommendation 30% Reduction in resource requirements

Confidential & Proprietary

18

Promise of Agentic AI in Decision Intelligence

Agentic AI is a type of artificial intelligence that operates
autonomously to achieve specific goals taking context-aware actions to
drive faster, smarter business decisions..

Key Characteristics · Autonomy: Acts independently to perform tasks
without constant direction. Agents can analyze data, evaluate options,
and make informed decisions, all while being able to self-optimize. ·
Goal-Oriented: Focuses on achieving specific objectives, devising plans
to reach them, and optimizing actions accordingly. · Interaction and
Adaptation to Changing Environments: Perceives and interacts with
dynamic environments, adapting strategies in real-time. · Continual
Learning and Improvement: agents employ reinforcement learning to
enhance performance over time.

Agentic AI Cycle

Make Decision s

Learn from Interactions

Take Actions

Confidential & Proprietary

Traditional vs Agentic AI

Agentic Al: A Comprehensive Framework for Autonomous Decision-Making
Systems in Artificial Intelligence

Key Performance Metrics of Agentic AI Systems

100

94.8

82.5 80

60

92.1 75.3

88.9 68.7

93.4 85.2

40

20

0

Decision

Response Time

Adaptation

Accuracy

Rate

System Reliability

Traditional AI (%)

Agentic AI (%)

Source: International Journal of Computer Engineering & Technology, 2025
Key Performance Metrics of Agentic AI Systems

19 1

Decision Agents

Decision Agents A Decision Agent are purpose-built to improve how
decisions are made. These agents are deployed to assist, augment, and/or
automate different steps in the process ­ from identifying decision
triggers, framing the problem and synthesizing information to generating
recommendations and when appropriate, executing actions.

Insight Agents "What's happening" Key Functions: Detect decision
triggers Understand questions within decision context Pull and connect
data sources Synthesize data and provide fit for purpose insights

Recommendation Agents "What should we do" Frame decision logic
Understand inter-relationship between insights Contextualize and
automate recommendations Act on recommendations

Confidential & Proprietary

Optimization Agents "It Gets Done ­ and improves over time" Driver
recommendation agent Monitor and evaluate outcomes Improve
recommendations through feedback loops 20

Let's take an example:

1 Decision Trigger

01

Analytically connecting business goals & KPI's to leading indicators
(metrics)

02

Alerts indicate situation has changed and impact on objectives

03

Decision is triggered and routed to owner

Turns

Gross Margin %

Inventory \$

COGS

SKU Productivity Market shar e Gross Margin Gr oss Margin (-2.4%)
\$VSolhuamr e (X.XMM lb (-3.0%) Price Gap (+0.5%) COGS X.X (+4.0 %) Pro
mo Investment X.X ( +0.1%) Consumer Sentiment X .X (+0.1) Pack S ize
(+0.5%) Source: Cloverpop

100 14 1 1 3 9 87

Beg. COGS Promo Pack Consumer Other End.

Period

Activity Size Sentiment

Pe r io d

Platform automatically identified: COGS increase in key ingredients \$
share loss in last 13 weeks Price gap has widened vs. competitors
Consumer sentiment is stable Margin fell below threshold

How Sh ould we re spo nd to commodity cost increase? Where can we r
educe cost to offset ma rgin impact? What promotional activities can we
d o to offset con sumer sentiment d rop? What promotional activities can
we d o to offset con sumer sentiment d rop?

Confidential & Proprietary

21

Let's take an example: Decision Model & Orchestration:

2

Decision Model & Orc hes trat ion

Decision: How should we respond to commodity price change?

What will be the financial impact? What is our COGS exposure?

Finance Partner

VP Brand

What will be the revenue impact from inc. price?

Key Functions: · What questions / sub-questions do we need to answer? ·
Which role / team / person should answer should be responsible? · Do we
have the data to answer the question?

How much cost should we pass onto consumers? If price change X, what
will be impact to Volume / Sales How will channel / customer response
look like?

D&A Team

Confidential & Proprietary

22

Let's take an example:

3-5

Decision Making & Learning

Data Sources ERP Fin. Data Warehouse Spend Cont. Tower Forecast Sim.
Price Elasticity Human input

Insight Agents

Recommendation Agents Decision: How should we respond to commodity price
change?

VP Brand

Optimization Agents

What will be the financial impact? Total margin impact estimated at
-\$5.6M without pricing actions; breakeven possible with targeted 4­6%
price adjustments What is our COGS exposure? 60% of COGS tied to three
volatile commodities; 12% YoY increase projected. What will be the
revenue impact from inc. price? Estimated +\$8.2M uplift, but 4% churn
risk in price-sensitive segments How much cost should we pass onto
consumers? Recommend passing on 50­60% of cost increase to maintain
competitiveness and limit volume erosion. If price change X, what will
be impact to Volume / Sales 3% volume decline projected if price
increases by 5% across core SKUs How will channel / customer response
look like? Mass retail may resist 5% hike; premium channel more
tolerant. Confidential & Proprietary

Fin anc e Partner D&A Team

Decision Insights Improve future decision models Improve Reco's through
feedback loops 23

Knowledge graph as the critical component

Goals

Market Share Stability/Growth

Business Decisions Business Issues

EFFICIENCY Do we need to optimize our media spend (reach, frequency,
allocation)?

Am I reaching enough con sumers (reach & frequen cy )?

Am I reaching the right con sumers efficiently? (allocation)

Drive Holistic Insights & Recommendations

Insights

Is my brand being "heard" by my target audience?

Are we spending in the right places

Derived Variables

media reach and frequen cy analysis

digital publishers' delivery/CTR analysis

impact by platform

awareness

Data Sources

media data

attribution study

MMM

A&U

Ad Tracking

Brand Equity

Enhance Impact by coCnonnfiedecnttiailn&gPromprieutalrtyiple data
sources

24

The data-to-decision disconnect

\$160bln spent on data, analytics and Ai...

Predictive analytics

Global capabilities

Reporting Factory

Prescriptive Analytics

Cloud

Advanced Forecasting

Data Marketplace

Data Lakes

Customer Data

Data Catalogue

ILLUSTRATIVE

HOW DO I USE THIS INFORMATION TO MAKE A DECISION?

Confidential & Proprietary

...Not making it to decisions

Where do we innovate? How do we grow?

Which supplier should we use? Do we pursue this acquisition?

Upwards of 60% of Insight and Analytics investments is wasted each year
Sou rce : 4i's 2 016 De cision Awarene ss Stu dy. 350 Execs (VP+) in 50
companies; Q uote ­ potential client interview 25

Barriers to taking effective decisions

Data-to-decision barrier manifests itself in three key ways

Data Quality, Access & Connectivity

Qualit y

Issues Acces s

Harmonizatio n

Data

Barriers to synthesis

X1 y1 X2 y2 X3

X7

y3

X9

Data to Decision Elevation Challenges

Decision Recom mendation Insi gh ts Data

"Now what" "S o-w h at" "W ha t"

Data quality, access and connectivity issues

Inability to connect multiple data sources to drive holistic
recommendations

Unsure of how to leverage data to drive recommendations (Now-what), not
just observations (what)

Confidential & Proprietary

26

Modern Data Ecosystem Unifying Data, AI, and Business Logic to Deliver
Actionable Growth Recommendations

Unified Data Lake

Purpose Built Data Products & models Customer

Knowledge

ERP

Adob e Analytics

CRM/ ...other Loyalty sources

Reviews

Product Price & Promo

· Gives a complete view of business performance, combining sales,
category trends, and vendor data. · Uncovers the "due-to" factors behind
revenue changes, like traffic drops, pricing shifts, or promo impact. ·
Contextual, curated information becomes knowledge. Ready for AI-driven
recommendations to improve assortment, pricing, promotions, or digital
shelf visibility.

Desired Outcomes

Governance Mechanisms & Value Based Roadmap Organization

Structure & Skillsets

Service Designs

DevOps, MLOps

Degree Of Federation

Value from Data

Persona Needs

Applications & Toolsets MDM

AI/ML Data Management &Go ve rn ance Data Platform

Data Products Data Privacy & Security Confidential & Proprietary

Insight-Driven Culture

Actions-to-Take delivered Merchant Add SKU X to match competitor B
premium category selection and prevent customer leakage. Adjust pricing
for SKU X from \$499 to \$479 to undercut Competitor X \$489 price while
maintaining margin. Inventory Offer X% discount on slow moving invenry
products SKU X, Y, and Z Marketing Build lifecycle campaigns for
previous buyers with complementary SKU suggestions (e.g., if they bought
subway tile, recommend coordinating trim or grout). 27

Realizing Full Value of Data

The goal of developing data products isn't to generate better data; it's
to generate value. The lion's share of the potential value to a company
generally comes from five to 15 data products

A data product is built from the beginning to be reused and extended to
meet a broad range of business cases. that are designed to collect,
organize, and manage data sets to be easily consumed by various teams or
systems

Data Suppliers

Structured, unstructured and semi-structured sources

CREATE TRANSFORM, HARMONIZE, AGGREGATE, STORE

PersUsoenra hip

Knoowf SleMdegse Gov

Data Products MeaVsauluraeble

DATA

ernance

Owners

Data Consumers

DATA MARKETPLACE BUSINESS INTELLIGENCE, JUPYTER NOTEBOOK, APIS, EXCEL,
DATABASES

INTERNAL Analysts Data Scientists Developers Business Users EXTERNAL Buy
or Request Partners

Confidential & Proprietary

McKinsey 28 2

Closing Data to Decision Divide

Knowledge with consideration and calculation becomes Wisdom Information
with context applied becomes Knowledge Data refined and structured
becomes Information Operational Systems

Wisdom Actions, decisions, future-looking

Knowledge

Contextual, curated, learnings

Information

Structured, organised, useful

Dai ta External Data Providers

Raw, unstructured, messy Coding Languages

Cloverpop, the leader in Decision Intelligence, and DataArt, the leader
in data and tech solutions, have partnered to bridge the
data-to-decision divide in supply chain.

Together, we combine cutting-edge AI-driven decisioning, scalable
cloud-native platforms, and deep domain expertise to transform
fragmented data into actionable and collaborative decisions -- helping
enterprises navigate volatility, optimize trade-offs, and act with
confidence at the speed of modern demand signals.

Confidential & Proprietary

29

Thank you! Following this webinar, you will receive a link to download
this webinar presentation! Confidential & Proprietary

Let's stay in touch! Lanny@Cloverpop.com Oleg.Royz@DataArt.com
jmehmet@aim.com Book a demo at www.Cloverpop.com 30