1 of 23

AI Agents and Workflow Automation

Presented by:

Arvind Sridharan

Head of Strategic Partnerships – Americas

04/14/2026

2 of 23

Rackspace as the operator of the full-stack from agents to infrastructure

With an Outcome-as-a-Service model

2

Edge to core to cloud

Core infrastructure

Cloud platforms

Forward Deployed Engineers (FDE)�& managed services

Cyber resilience

Cloud compliance

Core business applications

Developer-ready tools

Enterprise AI platforms

The question I always come back to when I look at this diagram: what does it actually take to go from a governed AI environment to real business outcomes?

Having the right infrastructure, the right data layer, the right decisioning environment gets you to the starting line — the work to get to business outcomes is what happens next

Enterprises need agents that act at the workflow level, at the interaction level, at the pace the business actually moves

And those agents need to work within the governed environment — not around it

That combination — governed infrastructure plus enterprise-grade agentic automation — is exactly what Rackspace and Uniphore have built together

Uniphore brings Business AI Cloud — orchestrating SLMs, LLMs, enterprise data, and deterministic agents into governed, end-to-end workflows

Open, sovereign, and composable — enterprises maintain full control over data, intellectual property, cost, and governance while scaling AI-driven operations

Purpose-built for regulated enterprises�
A month ago, we jointly launched what we believe is a new category: Agents-to-Infrastructure architecture — delivered as an outcomes-based service

Healthcare, financial services, insurance, public sector — enterprises in these industries can now move fast on AI without trading away the governance and control their business requires

Please welcome Umesh Sachdev, CEO and Founder of Uniphore, joining us virtually — Umesh, thank you for being here

3 of 23

About Uniphore�The Business AI Company

Business AI Cloud

Sovereign, Composable, & Secure AI Cloud with end-to-end AI Architecture to accelerate time-to-value for enterprises

Customers

Founded in 2008
HQ: Palo Alto, CA
Global Presence
~800 employees
Hypergrowth YoY

2 years in a row

Front Office

Growth

Context-Aware Inferencing Layer

Back Office

FSI | Tech | Insurance | Telecom | Energy

Solutions

KEY STAKEHOLDERS

Customers & End Clients

~2,000

Tech Leadership

Small Language Models

Automated fine-tuning

Agentic Process Discovery

AI-driven work mapping

Multi-Agent Orchestration

Deterministic agents

Context-Aware Inferencing

Domain awareness

Automated Data Prep

Enterprise-scale AI data

82 Patents | 36 Ph.Ds.

Research-Driven Acceleration

3

4 of 23

Business AI Cloud

A complete AI stack spanning agents, models, knowledge, and data

DATA LAYER

KNOWLEDGE LAYER

MODEL LAYER

AGENTIC LAYER

INFERENCING LAYER

KNOWLEDGE LAYER

MODEL LAYER

AGENTIC LAYER

4

5 of 23

Sovereign

Cloud Agnostic
Data Platform Agnostic
Model Agnostic
IT Ownership of Fine-tuned Models

Composable

Data Platform and Pipelines
Vector, Graph, Data Stores
Models, Tools, and Agents
APIs and Open Standards

Secure

Observability & Auditability
Secure Model Training
Differential Privacy
Data and Model Governance

Agentic Layer

Development, Deployment, and Orchestration of AI Agents

Model Layer

Model Garden and Orchestration

Knowledge Layer

Choice of Models to Fine-tune for Specific Business Domains

Data Layer

Data Acceleration for AI

Agentic Process Discovery

Ground-truth workflows to enrich domain-specific SLMs

Inferencing Layer

5

6 of 23

Outcome-as-a-Service

AI FULL STACK

PRE-PACKAGED INDUSTRY SOLUTIONS

DATA ENGINEERING

UNIPHORE

MANAGED SERVICES

INFERENCING SERVICES

INFERENCING OPTIMIZATION

DATA SERVICES

INFRASTRUCTURE SERVICES

FDE SERVICES

RACKSPACE

6

7 of 23

Industry’s first agents-to-infrastructure architecture to move AI from pilot to production at scale

One governed platform for data, models, agents and deployment- built for production
Domain-tuned models grounded in customer data, delivering explainable outputs while keeping IP proprietary
Intelligent agents that automate workflows across enterprise systems with human-in-the-loop governance
Unified data fabric connecting all sources into an AI-ready ontology with compliance-ready governance

Builds, operates and stays accountable for outcomes- wherever the workload runs
Infrastructure built for production AI: optimized compute, flexible scaling, predictable costs
Compliance-first architecture with data residency controls for GDPR, HIPAA and regulated sectors
Audit-ready from day one: model versioning, decision logging and compliance reporting
End-to-end support – we take the platform to market, deploy it, and operate it, guiding customers from consulting through day 2 operations

+

Rackspace takes the Uniphore platform to market, deploy it and run it.

QUESTIONS

Question 1: "Umesh, you have worked with enterprises across their AI adoption journey — across every industry, every size of organization. In your experience, what is actually blocking enterprises from getting AI into production?"

Question 2: "Uniphore describes Business AI Cloud as bridging the gap between consumer AI and enterprise AI. What does that distinction mean practically — and why does it matter for the customers we both serve?”

Question 3: "The partnership is structured as an outcomes-based service. That is a significant commercial commitment. What does that mean operationally — what does a customer actually experience?”

Question 4: "Earlier today the audience heard about Palantir's role in governing the AI operating environment. You operate at the agent layer. Help us explain how these two partnerships relate — and how Rackspace integrates both without creating complexity for the customer."

�Question 5: ”Our recent announcement describes a shared ambition to unlock $100 million in enterprise AI deployments. For this audience, ambition needs to come with structure. What are the target segments, and how will both companies measure progress?"

8 of 23

8

9 of 23

Business AI Cloud

Demo

9

10 of 23

10

11 of 23

Use-Case �Selection Criteria

11

12 of 23

A Framework for Disciplined Prioritization

Business Impact

Speed to Value

Ability to Evolve

Reusability

Integration Simplicity

Business Impact: Does it materially move a needle that is being tracked?

Speed to Value: Can we demonstrate measurable progress within a month to maintain momentum and sponsorship?

Integration Simplicity: Does it work with existing systems and clear owners to avoid architectural overreach?

Reusability: Does it create assets, data- pipelines, agent patterns—that accelerate the next initiative?

Ability to Evolve: Is there a clear 6–12-month roadmap beyond a Pilot?

12

13 of 23

Case Studies

14 of 23

Use Cases Driving Accelerated ROI

Domain-specific AI-powered solutions are transforming customer operations and accelerating returns through Uniphore’s solutions in partnership with leading consulting firms.

With AI-Powered Platform from Uniphore

Ticket Triage & Resolution

Business Problem

Contract IQ

Solution

Energy production faces millions in contract leakage
Improper use of equipment and unverified material spend
Overbilling and missed recovery opportunities

Autonomous Contract Risk Agents
Fraud & Compliance SLM
Contract IQ Knowledge Base
Unified Contract–Invoice Data Layer

Outcomes

Reduced leakage and overpayment

Higher fund recovery rates

Earlier detection of non-compliance and discrepancies

Business Problem

Billing Explanation & Diagnosis

Solution*

Massive complexity. > 9,000 possible billing scenarios.
Suboptimal transparency and CX erode customer trust.
Explainability accuracy is paramount, but agents aren’t enabled.

Domain-specific billing
SLM Billing knowledge base
Synthetic billing data pipeline

Outcomes*

Increase CSAT

Improve Agent Experience

Lower Average Handle Time

* Illustrative of POC and future phases

Business Problem

Solution*

Lack of context reuse slows time to resolution and constrains productivity
Reactive operating model increase escalation volume
Suboptimal metadata discipline requires manual tagging of tickets

Autonomous Support Triage Agents
Domain-Tuned SLM for Ticket Resolution
Customer Support Knowledge Base
Unified Case & Technical Data Layer

Outcomes*

Reduce Cycle Time

Higher Operational Efficiency

Higher First-Touch Resolution

* Illustrative of POC and future phases

15 of 23

Overview of Contract IQ

Energy production faces millions in contract leakage

Improper use of equipment and unverified material spend

Overbilling and missed recovery opportunities

THE BUSINESS PROBLEM

Connected & Enriched Contract Documents, Clauses, and Invoices

Established a Contract IQ Knowledge Base and Fine-Tuned �Small-Language Model

Operationalized Domain-Specific SLM with Expertise in Contract Leakage, Fraud Detection, & Compliance

Deployed AI Agents to Calculate Contract Leakage, Detect Non-Compliance, & Flag Invoice Discrepancies

Reduced leakage and overpayment

Higher fund recovery rates

Earlier detection of non-compliance and discrepancies

16 of 23

Overview of Billing Explanation & Diagnosis

Massive complexity. > 9,000 possible billing scenarios.

Suboptimal transparency and CX erode customer trust.

Explainability accuracy is paramount, but agents aren’t enabled.

THE BUSINESS PROBLEM

Connect to Synthetic Billing Data

Establish a Billing Knowledge Base and Fine-Tuned �Small-Language Model

Operationalize Domain-Specific SLM with Expertise in Billing Scenarios, Explanations & Standard Operating Procedures

Not in Scope

Increase CSAT

Improve agent experience

Lower average handle time

Pre-deployment | Illustrative of POC and future phases

17 of 23

Overview of Ticket Triage & Resolution

Lack of context reuse slows time to resolution and constrains productivity

Reactive operating model increase escalation volume

Suboptimal metadata discipline requires manual tagging of tickets

THE BUSINESS PROBLEM

Connect to Historical Case Notes, Resolution Docs, and Technical FAQs

Establish a Customer Support Knowledge Base and Fine-Tuned �Small-Language Model

Operationalized Domain-Specific SLM with Expertise in Ticket Triage and Resolution, Historical Case Procedures, and Technical Specs

Deploy AI Agents to Enrich Ticket Metadata, Route Tickets, Recommended Resolution Path, and Escalate Tickets Proactively

Reduce cycle time

Higher operational efficiency

Reduction in handoffs (higher first-touch resolution)

Pre-deployment | Illustrative of POC and future phases

18 of 23

Expanding the Partnership

17K

Customer Calls

Directly from Farmers

2 Fine Tuned SLMs

BillExplainer

Sales-Retention

Conversational Insight Agent

Deployed Insights Agent to analyze customer interactions
Surfaced actionable insights to improve experience and �reduce churn

UNIPHORE

+

FARMERS

Real-Time Guidance Agent

Agent Assist live in Customer Service to guide agents in real-time
Improved AHT, accuracy, and consistency

AI Agency

Fully AI automated agency incl. marketing, sales, servicing etc.
Drive up-sells and renewals for “Orphaned accounts”
Connects customer and policy data for personalized upsells

Retention

Discover analyzes interactions to predict churn.
Agentic actions trigger personalized retention outreach.

19 of 23

19

QUESTIONS

Question 1: "Umesh, you have worked with enterprises across their AI adoption journey — across every industry, every size of organization. In your experience, what is actually blocking enterprises from getting AI into production?"

Question 2: "Uniphore describes Business AI Cloud as bridging the gap between consumer AI and enterprise AI. What does that distinction mean practically — and why does it matter for the customers we both serve?”

Question 3: "The partnership is structured as an outcomes-based service. That is a significant commercial commitment. What does that mean operationally — what does a customer actually experience?”

Question 4: "Earlier today the audience heard about Palantir's role in governing the AI operating environment. You operate at the agent layer. Help us explain how these two partnerships relate — and how Rackspace integrates both without creating complexity for the customer."

�Question 5: ”Our recent announcement describes a shared ambition to unlock $100 million in enterprise AI deployments. For this audience, ambition needs to come with structure. What are the target segments, and how will both companies measure progress?"

20 of 23

Backup Slides

21 of 23

AGENTIC LAYER

Agentic Workflows

MCP Based Orchestration

3^rd Party

Agents

Natural

Language Inputs

Industry

Agents

Data

Agents

Data

Automation

Data Transformation

Data�Acceleration

Data�Fabric

MODEL LAYER

Model

Orchestration

Guardrails

Model

Garden

Industry�Models

BYO�Models

KNOWLEDGE LAYER

RAG

Knowledge

Graphs

Enrichment

Fine-Tuning

SLM

Factory

DATA LAYER

Business AI Cloud

DATA LAYER

KNOWLEDGE LAYER

MODEL LAYER

AGENTIC LAYER

INFERENCING LAYER

KNOWLEDGE LAYER

MODEL LAYER

AGENTIC LAYER

INFERENCING LAYER (CAIO)

Multi compute Dynamic routing

Auto Scaling

Observability & Governance

Guardrails & Model Security

Cost

Optimization

21

<Directions>�Tailor the full-stack narrative to the buyer’s maturity:
For CIOs: emphasize control, security, and integration.
For business units: highlight speed to value and use-case agility.
---

This is the full power of the Uniphore Business AI Cloud.
While others are stitching together tools, plugins, and point solutions, we’ve built a unified platform from the ground up—across data, knowledge, models, and agents.

You don’t need to be locked into our ecosystem in order to use any of the layers on their own.

Each layer is designed to do the heavy lifting—turning fragmented enterprise inputs into intelligent, orchestrated outcomes.
This isn’t just GenAI bolted onto your business. It’s AI that’s embedded in how your business actually runs.
Backed by a powerful platform that’s designed to help you deliver real outcomes with AI that's ready to work right now, without compromising.

22 of 23

Selecting the Right PILOT Use Case

GUIDELINES & BEST PRACTICES

Material impact on cost, revenue, risk, customer experience, or operational efficiency Solving company-level priorities or problems get leadership support and accelerate momentum	Select use cases which solve significant business pain and are highly visible Focus on use cases which can show measurable improvement in the first 60-90 days from deployment Confirm that leadership views this use case as a priority
Early wins build momentum and executive confidence. Quick results demonstrate ROI, validate capabilities, and secure continued investment.	Prioritize processes with clear KPIs and visible pain-points that can be automated or enhanced with minimal integration. Focus on use cases where success is measurable and transferrable.
During a PILOT, velocity matters. Use cases that require less effort to implement accelerate deployment and minimize organizational friction.	Choose use cases with accessible data, defined ownership, and established process understanding. Prioritize practical application over novelty or experimentation.
AI that can’t scale can’t sustain value. Reusable components reduce OPEX, shorten deployment cycles, and create organizational muscle memory.	Leverage modular, composable components that can be redeployed across teams and domains. Capture learnings, templates, and patterns in a shared AI playbook to accelerate subsequent efforts. (Uniphore to lead this enablement).
Strong POCs expand over time. Ideal use cases allow for additional data, integrations, and automation once in production.	Define a 6–12 month roadmap from proving value to scaling value to cross-functional orchestration. Anchor the roadmap to measurable business outcomes.

Dimension

Why It’s Important

Considerations

SPEED TO VALUE

INTEGRATION SIMPLICITY

REUSABILITY

ABILITY TO EVOLVE

BUSINESS VALUE

22

23 of 23

Rackspace + Uniphore

Accelerating enterprise AI adoption

What we bring
1	Multi-Environment Deployment Public, Private, hybrid, on-premise
2	Infrastructure Compliant, Secure, Sovereign
3	Technology expertise Palantir-certified Forward Deployed Engineers

Platform power
1	Business AI Cloud AI platform integrates data ingestion, model training, agent orchestration, & deployment
2	SLM Architecture Domain focused Fine-tuned SLM’s
3	Data & Knowledge Layers Unified data fabric with Zero copy and Zero ETL

As a certified partner of Uniphore, Rackspace, we are delivering the industry’s first Infrastructure to Agents architecture as an outcome-based service

Together, we solve the complex data and tech debt challenges preventing you from realizing AI ROI and driving data-driven outcomes across your enterprise.
Platform expertise Uniphore-certified Forward Deployment engineers building with precision to unlock full platform performance
Rapid delivery Solutions deployed in Weeks No need for long LLM training cycles.
Full lifecycle support From AI advisory through production, ongoing optimization, and Day 2 Services.

By combining Uniphore's platform with Rackspace’s governed cloud operations, we are becoming the de facto choice for enterprise AI at scale.

23