# Anthropic and DXC Bring Claude into Regulated Enterprise Systems: Why AI Workflows Need Cost Control
Anthropic has announced a multi-year global alliance with DXC Technology. DXC is one of the world’s largest IT services companies and operates systems used by banks, airlines, insurers, manufacturers and government agencies.
According to Anthropic, DXC will train tens of thousands of Claude-certified forward-deployed engineers. These engineers will work directly inside customer organizations to bring Claude into the systems that regulated industries depend on.
This update matters because it shows that enterprise AI is moving beyond chat.
In consumer AI, users often ask a question in a chat box and get an answer. In enterprise AI, especially in regulated industries, AI must work inside real systems with security, compliance, auditability and human review.
Enterprise AI is moving into core business systems
A simple AI chatbot is not enough for industries such as banking, aviation, insurance, manufacturing and government.
These organizations rely on systems that handle transactions, claims, operations, maintenance and customer support. These systems often have strict requirements for reliability, privacy, compliance and accountability.
When AI enters these environments, the goal is not only to generate fluent answers. The system must support:
- secure access
- controlled workflows
- human review
- audit trails
- predictable cost
- model selection
- task-level monitoring
This is why AI Workspace products matter. A useful workspace is not just a chat interface. It should help users manage models, context, cost, prompts and outputs.
DXC OASIS shows how agentic workflows are being deployed
Anthropic says DXC has already used Claude internally before bringing it to clients. One example is DXC OASIS, an AI-native orchestration platform for managed services.
Claude is now the default foundation model powering the platform’s agentic workflows. DXC also says Claude helped generate more than 95% of the code for OASIS, with software engineers reviewing the output.
This shows a practical model for enterprise AI:
AI does not replace professionals completely. It handles routine work, analysis, generation and workflow steps, while humans review important results and decisions.
That is a more realistic path for adoption in regulated industries.
Regulated industries need better model selection
Enterprise AI cannot use one model for every task.
Different tasks require different model choices:
- low-cost models for simple classification
- long-context models for document analysis
- coding models for modernization and refactoring
- security-focused models for cyber workflows
- stronger reasoning models for high-value decisions
- human review for regulated or high-risk outputs
This means model selection is becoming a core product capability.
Users do not only need a dropdown list of models. They need to understand which model fits the task, how much it may cost and when a stronger model is worth using.
Toket AI is built around this problem through Token Calculator, Prompt Optimizer and AI Workspace.
Token cost becomes an operating cost
When AI becomes part of enterprise systems, token cost becomes an operating cost.
A workflow in banking or insurance may include:
1. reading customer records or tickets 2. retrieving policies or historical context 3. classifying the task 4. generating a recommendation 5. sending the result to human review 6. recording the final decision 7. generating follow-up actions
This is not one model call. It is a multi-step workflow.
If an organization runs thousands of these tasks per day, token usage can grow quickly. Long context, code analysis, security review and agentic workflows can make cost even more unpredictable.
A Token Calculator helps users estimate the cost of the full workflow, not just the price of one API call.
Prompt Optimizer reduces waste in enterprise workflows
Prompt quality becomes more important as workflows become more complex.
A vague instruction can cause the model to:
- produce unstable formats
- ignore business rules
- mix up sources
- miss escalation conditions
- generate overly long answers
- require repeated retries
All of this increases token usage.
A stronger enterprise prompt should define:
- role
- task goal
- allowed sources
- output format
- risk level
- escalation rules
- human review requirements
- uncertainty handling
Prompt Optimizer helps turn vague requests into structured workflow instructions. Better prompts reduce retries, improve output quality and lower wasted tokens.
AI Workspace should make the process visible
Enterprise users need more than a final answer.
They need to see:
- current task stage
- current model
- context length
- token or credit usage
- source materials
- outputs that need review
- decisions made by humans
- results that should be saved
This is the difference between a chat tool and an AI Workspace.
A chat tool answers a question. An AI Workspace helps users manage a task.
The Anthropic and DXC alliance shows where enterprise AI is going: models will be embedded into real workflows, and users will need better tools to manage cost, context and reliability.
What Toket AI users should take away
Anthropic’s alliance with DXC shows that the next stage of AI adoption is not only about stronger models.
It is about bringing models into real business systems in a controlled way.
For Toket AI users, the practical takeaways are:
1. Choose models based on task type, not only brand name. 2. Estimate token cost before running long workflows. 3. Use Prompt Optimizer to define clear roles, boundaries and output formats. 4. Use AI Workspace for multi-step tasks instead of a single chat box. 5. Keep human review for regulated, high-risk or business-critical outputs.
Enterprise AI will not be won by the smartest model alone. It will be won by better model selection, better prompts, better cost control and better workflow management.