Business 12 min read
How Software & Technology Organizations Are Adopting and Budgeting for AI Technologies
Artificial intelligence has moved from the innovation lab into the core business strategy of software and technology organizations. A few years ago, companies were primarily experimenting with large language models, chatbots, and generative AI. Today, organizations are building AI into software development, customer service, sales, finance, operations, cybersecurity, knowledge management, and decision-making.
The major change is not simply that companies are using more AI. They are also changing how they budget for, govern, deploy, and measure AI.
The central question for technology leaders has evolved from:
“What can AI do?”
to:
“Where should we deploy AI, how much should we invest, and what measurable business value will it create?”
1. AI Has Become a Strategic Technology Investment
Enterprise AI adoption is accelerating across industries. A 2026 study analyzing ChatGPT Enterprise usage across more than 1,500 organizations and 17 million messages found that enterprise AI usage has grown rapidly, with adoption spanning technical work, writing, communication, and information synthesis. Adoption is particularly concentrated among larger organizations with significant R&D and SG&A activity.
This is important because AI is no longer restricted to data-science teams.
A software engineer may use AI for:
Code generation
Code review
Test creation
Debugging
Documentation
Architecture exploration
A salesperson may use it for:
Account research
Meeting preparation
Proposal generation
CRM updates
Sales-call analysis
A manager may use it for:
Meeting summaries
Decision tracking
Project-risk analysis
Business reporting
Knowledge discovery
A customer-support team may use AI agents to:
Understand customer questions
Search knowledge bases
Resolve simple issues
Escalate complex cases
Update support systems
Therefore, organizations increasingly view AI as a horizontal capability rather than a single department's technology.
2. The Four Stages of Enterprise AI Adoption
Most organizations do not move directly from experimentation to enterprise-wide AI.
They typically progress through four stages.
Stage 1: Experimentation
Employees and small teams begin experimenting with ChatGPT, Claude, Gemini, coding assistants, open-source models, and AI APIs.
The initial budget may be relatively small.
For example:
$10K–$50K for tools
Cloud/API consumption
Small proof-of-concepts
Developer experimentation
The primary question is:
“Can AI solve this problem?”
Stage 2: Department-Level Adoption
Once a successful use case is identified, organizations deploy AI within specific departments.
Examples include:
AI coding assistants for engineering
AI assistants for sales
AI search for legal teams
AI document processing for finance
AI support agents for customer service
Now the budget begins moving from an innovation budget toward the departmental operating budget.
The question becomes:
“Can we demonstrate measurable productivity or business improvement?”
Stage 3: Enterprise Scaling
Successful use cases are integrated with enterprise systems.
For example:
Employee
↓
AI Assistant
↓
Enterprise Knowledge
↓
RAG / Search
↓
AI Agent
↓
MCP / APIs / Tools
↓
CRM / ERP / Jira / HR / Finance
↓
Business Action At this stage, organizations need much more than an LLM.
They need:
AI platform architecture
Identity and access management
Data governance
RAG
Vector databases
Model gateways
Agent orchestration
Evaluation frameworks
Observability
Security
Guardrails
Human approval
Auditability
The question becomes:
“How do we safely scale AI across the organization?”
Stage 4: AI-Native Organization
The most advanced organizations begin redesigning business processes around AI.
Instead of asking:
“Where can we add AI to the existing process?”
they ask:
“If AI were available from day one, how would we design this process?”
This creates AI-native workflows where agents perform portions of the work traditionally handled manually.
3. What Organizations Are Actually Budgeting For
AI budgets are much broader than simply buying an LLM subscription.
A mature enterprise AI budget can include at least eight categories.
1. AI Models
Organizations may consume:
OpenAI models
Anthropic models
Google models
Microsoft models
Open-source models
Specialized models
Companies increasingly use multiple models rather than depending on a single provider.
For example:
Simple task → Low-cost model
Complex reasoning → Premium model
Private workload → Open-source model
Sensitive data → Controlled/private deployment This creates a new requirement: model routing and model governance.
2. Cloud and GPU Infrastructure
AI workloads require significant compute.
Organizations may spend on:
GPUs
AI accelerators
Cloud inference
Model hosting
Storage
Networking
Data centers
This is particularly significant for technology companies developing their own AI products.
The current AI infrastructure race is enormous. Reuters reported in August 2026 that investors are increasingly focused not only on how much Big Tech is spending on AI infrastructure but also on which companies can ultimately generate returns from those investments.
3. Data Infrastructure
AI is only as useful as the data available to it.
Companies therefore invest in:
Data lakes
Data warehouses
Vector databases
Knowledge graphs
Metadata systems
Data pipelines
Document processing
Enterprise search
This is where RAG becomes important.
A company may have thousands of:
PDFs
Emails
Meeting transcripts
CRM records
Jira tickets
Contracts
Policies
Product documents
An enterprise AI assistant needs controlled access to this information.
4. AI Agents Are Changing the Budget Equation
One of the biggest shifts happening now is the transition from AI assistants to AI agents.
A traditional AI assistant might answer:
“What is the status of Project X?”
An agent could potentially:
Search Jira
Retrieve project documentation
Analyze recent meetings
Identify overdue tasks
Check customer complaints
Assess risks
Prepare a report
Update the project-management system
Notify the project manager
This is much closer to workflow automation than conventional chatbot usage.
Organizations are therefore beginning to budget for:
AI + Data + Tools + Workflow + Governance
rather than simply:
AI + Chatbot
5. Real Industry Example: JPMorgan Chase
JPMorgan is one of the clearest examples of an organization treating AI as an enterprise technology capability.
In its 2025 annual reporting, JPMorgan said more than 90% of its engineers use AI coding assistants, while more than 65,000 Corporate & Investment Bank employees actively use its LLM Suite.
The bank is using AI for activities including:
Transaction screening
Software engineering
Client advice
Research
Meeting preparation
Market analysis
Risk management
Treasury forecasting
One particularly interesting example is transaction screening.
JPMorgan reported that AI enabled the organization to review more than twice the volume while reducing manual operator checks by half.
This illustrates an important principle:
AI investment is increasingly justified through operational metrics.
Instead of saying:
“We invested in AI because AI is strategically important.”
the organization can say:
“AI allows us to process twice the workload with fewer manual checks.”
That is a much stronger business case.
6. Real Industry Example: Morgan Stanley
Morgan Stanley provides another excellent example, particularly for enterprise knowledge management.
The company worked with OpenAI to build AI solutions for financial advisors. Its internal AI @ Morgan Stanley Assistant helps advisors retrieve information from the organization's knowledge base.
According to OpenAI's published case study, more than 98% of advisor teams actively use the AI assistant. Document access increased from approximately 20% to 80%, while AI helped advisors spend more time on client relationships.
But the most interesting part is not the chatbot.
It is the evaluation framework Morgan Stanley built around AI.
Before deploying AI use cases, the company evaluates:
Accuracy
Reliability
Quality
Real-world performance
Expert feedback
This demonstrates a critical change in enterprise AI:
AI evaluation is becoming part of the AI budget.
Organizations cannot simply purchase a model and assume that it will produce reliable results.
7. Real Industry Example: Accenture
Accenture demonstrates a different model.
Instead of merely adopting AI internally, it is building AI capabilities as a major business opportunity.
Accenture reported that in fiscal 2025 its revenue from generative AI and increasingly agentic AI reached $2.7 billion, roughly triple the previous year's level. Its generative AI bookings nearly doubled to $5.9 billion.
Accenture also invested heavily in:
Acquisitions
R&D
AI platforms
Training
Learning and development
It reported approximately $1 billion in learning and development investment during fiscal 2025, with employees completing approximately 47 million hours of training, with generative AI a major focus.
This illustrates another important trend:
AI budgets increasingly include people and skills.
Buying AI technology without training employees does not guarantee adoption.
8. AI Budgeting Is Becoming a Portfolio Exercise
Technology leaders increasingly evaluate AI investments as a portfolio.
For example:
InvestmentObjectiveMeasurementCoding AIDeveloper productivityEngineering hours savedCustomer AI agentReduce support costCost per ticketSales AIImprove sales productivityRevenue / conversionEnterprise RAGReduce information searchSearch timeMeeting intelligenceImprove executionAction completionAI forecastingImprove planningForecast accuracyAI automationReduce manual workHours eliminatedAI securityReduce riskIncidents prevented
This changes how CIOs and CTOs think about AI.
The question is no longer:
“How much does this AI tool cost?”
It becomes:
“What economic value does this AI capability generate?”
9. AI Budgeting Is Moving Toward ROI
This is arguably the biggest challenge facing enterprise AI.
Organizations have demonstrated that AI can generate impressive results in individual tasks.
But scaling those results across an enterprise is difficult.
A recent 2026 research paper describes this as a “deployment wall”: organizations can spend heavily on AI pilots without necessarily translating those pilots into measurable P&L impact. The research argues that production deployment friction—not simply model capability—is a major reason AI projects fail to generate value.
This is why AI budgets increasingly include:
Integration
Data preparation
Security
Governance
Evaluation
Monitoring
Change management
Workflow redesign
The cost of productionizing AI can be much greater than the cost of the initial proof of concept.
10. The Hidden AI Budget
Many companies initially underestimate the hidden costs.
Suppose a company wants to build an AI customer-support agent.
The obvious budget might be:
LLM API + application development
But the actual budget may look more like:
LLM
+
Cloud
+
Data
+
RAG
+
CRM integration
+
Security
+
Identity
+
Evaluation
+
Monitoring
+
Guardrails
+
Human escalation
+
Support
+
Training This is why enterprise AI projects can become significantly more expensive when they move from prototype to production.
11. Organizations Are Also Budgeting for AI Governance
AI introduces new risks.
Companies need to consider:
Hallucinations
Data leakage
Privacy
Bias
Intellectual property
Regulatory compliance
Unauthorized AI agents
Prompt injection
Excessive agent permissions
Incorrect automated decisions
For example, an AI agent that can read CRM data is relatively low risk.
An agent that can:
Modify customer records
Approve payments
Delete data
Send contracts
Change pricing
has a much higher risk profile.
Therefore, enterprise AI architecture increasingly looks like:
AI Governance
│
┌──────────────┼──────────────┐
↓ ↓ ↓
Security Evals Observability
│ │ │
└──────────────┼──────────────┘
↓
AI Platform
↓
Models / Agents / RAG
↓
Enterprise Systems 12. AI Spending Is Also Creating a Skills Budget
AI adoption requires new roles.
Organizations increasingly need:
AI Engineers
ML Engineers
AI Architects
Data Engineers
AI Product Managers
AI Platform Engineers
AI Security Engineers
AI Governance specialists
AI Evaluation specialists
Prompt/Context engineers
Agentic AI developers
But there is another important trend:
Existing employees are also being trained to use AI.
This is why companies such as Accenture are investing heavily in learning and development.
The future workforce will probably not be divided simply into:
AI employees vs. non-AI employees.
Instead, many roles will become:
AI-enabled roles.
13. The Software Engineering Function Is Being Completely Reshaped
Software engineering is one of the areas receiving the most AI investment.
The traditional development process:
Requirement
↓
Design
↓
Coding
↓
Testing
↓
Code Review
↓
Deployment is increasingly becoming:
Requirement
↓
AI-assisted Design
↓
AI-assisted Coding
↓
AI-generated Tests
↓
AI Code Review
↓
Automated Security Analysis
↓
CI/CD
↓
Human Approval JPMorgan's experience is a strong example: more than 90% of its engineers now use AI coding assistants.
This means companies may increasingly budget for developer productivity platforms rather than just individual coding assistants.
14. AI Infrastructure Is Becoming a Board-Level Investment
For hyperscalers and major technology companies, the scale is much larger.
They are investing billions in:
Data centers
GPUs
Networking
Power
Cooling
AI accelerators
Cloud infrastructure
This is different from an enterprise buying an AI SaaS product.
There are essentially two AI economies developing:
AI consumers
Companies using AI to improve their businesses.
Examples:
Banks
Retailers
Manufacturers
Healthcare companies
Software companies
AI infrastructure providers
Companies building the infrastructure that makes AI possible.
Examples include:
Cloud providers
Semiconductor companies
GPU manufacturers
Data-center operators
Networking companies
Current investment activity shows just how large this infrastructure layer has become.
15. AI Budgeting Is Moving From CapEx to a Hybrid Model
Traditional IT investment often separates:
CapEx
from
OpEx.
AI complicates this.
A company may have:
CapEx
GPUs
Data centers
Servers
Networking
OpEx
LLM API usage
Cloud inference
SaaS AI tools
AI agents
Data services
People investment
AI engineers
Consultants
Training
Research
Therefore, CFOs need a much more sophisticated AI cost model.
16. The Rise of FinOps for AI
AI also introduces the need for AI FinOps.
Consider an AI agent that processes millions of documents.
The company needs to understand:
Tokens consumed
Model cost
Retrieval cost
GPU utilization
API calls
Agent execution cost
Storage
Embedding cost
For example:
User request
↓
Agent
↓
5 tool calls
↓
3 retrieval operations
↓
2 LLM calls
↓
Final response The organization needs to know:
What did that single business task cost?
Eventually, AI cost accounting may become as normal as cloud cost accounting.
17. The Most Important Change: AI Is Becoming Outcome-Based
The strongest organizations are moving toward:
Budget → Capability → Outcome
rather than:
Budget → Tool
For example:
Weak AI business case
“We need $500,000 for an enterprise AI platform.”
Strong AI business case
“We will automate 40% of customer-support interactions, reduce average handling time by 30%, and save $1.5 million annually.”
The second argument is much easier for a CFO to approve.
18. What Should an Enterprise AI Budget Look Like?
There is no universal percentage that every company should allocate to AI.
Instead, organizations should build the budget around their maturity and objectives.
A practical framework is:
10–15% — Experimentation
Fund:
POCs
Model evaluation
New tools
Emerging technologies
20–30% — AI Platform
Fund:
Model access
RAG
Agent frameworks
APIs
Infrastructure
Data
15–25% — Production Applications
Fund:
Customer agents
Sales AI
Engineering AI
Finance AI
Operations AI
10–15% — Security & Governance
Fund:
AI security
Evaluations
Guardrails
Compliance
Monitoring
10–20% — People & Training
Fund:
AI engineers
Architecture
Training
Change management
These are planning ranges, not industry benchmarks. The appropriate allocation depends heavily on whether the organization is an AI consumer, AI software vendor, or AI infrastructure provider.
Conclusion
Enterprise AI is entering a new phase.
The first wave was about experimentation.
The second wave was about productivity.
The current wave is about production, automation, agents and measurable ROI.
Organizations are increasingly budgeting not only for AI models but for the complete ecosystem required to make AI work at enterprise scale:
Models + Data + Infrastructure + Agents + Integration + Security + Governance + People + Measurement.
The examples of JPMorgan, Morgan Stanley and Accenture demonstrate three different dimensions of this transformation.
JPMorgan is using AI to improve productivity and financial operations. Morgan Stanley is using AI to transform knowledge access and advisor workflows. Accenture is treating generative and agentic AI as both an internal capability and a major growth business.
The most important lesson is this:
AI spending is no longer about buying AI. It is about building an organization capable of turning AI into measurable business value.
And that is likely to be the defining enterprise technology challenge of the next several years.
