Business intelligence is changing quickly. Traditionally, creating an enterprise report meant understanding the data model, selecting dimensions and measures, building queries, designing visualizations, testing prompts, validating the output, and finally publishing the report.
With IBM Cognos Analytics 12.1.3, IBM is moving toward a different experience: combining governed business intelligence with agentic AI.
The latest release brings three particularly interesting capabilities into focus:
- Report Authoring Agent
- Model Context Protocol (MCP) integration
- Bring Your Own LLM (BYO-LLM)
Together, these capabilities show how Cognos is evolving from a traditional BI platform into a platform that can participate in broader AI-powered analytics workflows.
What is new in Cognos Analytics 12.1.3?
Cognos Analytics 12.1.3 continues the AI capabilities introduced in earlier 12.1 releases.
The major change is that AI is no longer limited to simply helping users find or summarize information. The Report Authoring Agent is now generally available, allowing users to describe what they want in natural language and have Cognos assist with creating or modifying reports.
For example, instead of manually starting with a blank report, a user could describe a requirement such as:
Create a report showing sales by region and product category, with a monthly trend and a filter for fiscal year.
The Authoring Agent can interpret the request and work with the governed Cognos environment to create the report. Existing reports can also be modified through conversational instructions, such as changing a visualization, adding a filter, or modifying a calculation.
This does not eliminate the need for BI knowledge. Instead, it changes how users interact with the BI platform.
Report Authoring Agent: From Requirement to Report
One of the most interesting developments in 12.1.3 is the Report Authoring Agent.
Traditionally, a report developer may receive a requirement like:
- Show total revenue
- Break it down by region
- Add monthly trends
- Add a fiscal-year prompt
- Highlight high-performing regions
The developer then translates those business requirements into Cognos report objects, queries, calculations, prompts, and visualizations.
With the Authoring Agent, natural language becomes another way to communicate those requirements.
The agent can help select appropriate measures and dimensions, create report layouts, apply visualizations, and work with existing reports. IBM states that the 12.1.3 Authoring Agent supports more report types and capabilities than its earlier preview, including chart styling, editing, duplication, prompt support, and multilingual authoring.
Why is this important?
It can reduce the amount of repetitive work involved in report development.
For example:
Traditional approach
Business requirement → Developer → Report specification → Development → Testing → Changes → Final report
AI-assisted approach
Business requirement → Natural-language instruction → AI-assisted report creation → Developer validation → Final report
The second approach does not remove development and testing. It can, however, shorten the distance between a business requirement and an initial report.
Does this mean Cognos Developers are no longer needed?
Not at all.
In fact, the role of a Cognos developer may become more focused on data understanding, governance, validation, performance, and solution design.
AI can help create a report, but a developer still needs to ask important questions:
- Is the correct data source being used?
- Are the measures calculated correctly?
- Are relationships between objects correct?
- Are filters behaving as expected?
- Does the report perform well with production volumes?
- Are security rules being respected?
- Does the result actually satisfy the business requirement?
- Is the visualization appropriate for the business question?
These are not simply report-authoring questions. They require understanding of the underlying data and business logic.
This is why I see AI-assisted Cognos development less as “AI replacing developers” and more as “AI changing what developers spend their time on.”
What is MCP?
Another major addition in Cognos Analytics 12.1.3 is the Cognos Analytics MCP Server.
MCP stands for Model Context Protocol.
At a high level, MCP provides a standardized way for AI assistants and applications to interact with external tools and systems.
In the Cognos context, this means Cognos can become a governed analytics tool that can be called by an MCP-compatible AI assistant or workflow.
IBM states that the Cognos MCP Server exposes more than 70 tools across areas including content operations, report execution, report updates, AI analysis, security, governance, and administration.
- Why MCP Matters for Enterprise BI
Imagine an organization has an AI assistant used by business teams.
A user asks:
“Show me the latest operational risk report and summarize the major changes.”
Instead of the AI system operating independently of the BI platform, an MCP-enabled workflow can potentially interact with Cognos to:
- Find the relevant Cognos content
- Execute the appropriate report
- Retrieve the result
- Analyze or summarize the information
- Return the insight to the user
The important part is that Cognos remains part of the governed analytics environment.
This creates an interesting architecture:
User → AI Assistant → MCP → Cognos → Governed Data → Report/Insight → AI Assistant → User
Rather than building another disconnected analytics layer, organizations can connect AI workflows to their existing BI environment.
- MCP and the Future of AI Workflows
MCP becomes particularly interesting when organizations start building larger AI workflows.
For example:
AI-powered Risk Management Workflow
User asks:
“Which risk areas have deteriorated this month?”
The workflow could potentially:
- Access governed Cognos content
- Run the relevant report
- Analyze the results
- Identify significant changes
- Summarize the findings
- Present the result to the user
The same concept could apply to finance, procurement, sales, HR, audit, compliance, and operational reporting.
IBM describes MCP integration as a way to connect Cognos with broader enterprise AI workflows while maintaining Cognos security and governance.
What is BYO-LLM?
The third major capability is Bring Your Own LLM, commonly referred to as BYO-LLM.
LLM means Large Language Model.
Previously, organizations using AI capabilities often had fewer choices regarding the underlying model.
With BYO-LLM in Cognos Analytics 12.1.3, organizations can choose the language model that fits their requirements instead of being restricted to a single model.
IBM’s documentation notes that Cognos 12.1.3 and later can use different LLMs, including models from different providers. For Cognos Analytics Certified Containers with Agentic AI, IBM also documents BYO-LLM support through a LiteLLM proxy integration.
Why Would an Organization Want Its Own LLM?
Different organizations have different requirements.
For example:
Data and privacy
Some organizations may have strict policies around where enterprise information can be processed.
Model selection
Different LLMs may perform differently depending on the task.
Cost
Organizations may want to evaluate the cost of using different models at scale.
Existing AI strategy
A company may already have an approved AI platform or model ecosystem.
Deployment requirements
Some organizations may prefer particular deployment patterns based on their security and infrastructure requirements.
BYO-LLM gives organizations greater flexibility while allowing Cognos governance to remain part of the analytics experience.
AI Does Not Remove the Importance of Governance
This may actually be one of the most important points in the entire release.
Enterprise AI cannot simply be about generating answers.
Organizations need to know:
Where did the data come from?
Who is allowed to access it?
Which metrics were used?
Can the result be trusted?
Can the activity be governed and audited?
Cognos is designed around governed BI, including certified data models, security controls, and controlled access to enterprise information. IBM positions the 12.1.3 release as combining agentic AI with governed BI rather than creating a completely separate AI analytics environment.
That distinction is important.
The future of enterprise AI is unlikely to be simply:
AI + Data
It is more likely to be:
AI + Trusted Data + Governance + Security + Business Context
What Does This Mean for Cognos Developers?
For Cognos developers, the skill set is likely to evolve.
Traditional skills remain important:
- Report development
- Data modeling
- SQL
- Calculations
- Dashboard development
- Prompt design
- Performance tuning
- Security
- Testing
- Troubleshooting
But new skills are becoming increasingly valuable:
- Generative AI
- Prompt engineering
- Agentic AI
- LLM concepts
- AI governance
- MCP
- API-based integrations
- Data governance
- AI-assisted development
A Cognos developer who understands both BI fundamentals and AI-enabled workflows can become much more valuable than someone who only knows how to build a traditional report.
Will AI Replace Traditional Cognos Reporting?
I don’t think the answer is simply yes or no.
Simple reporting tasks are increasingly becoming candidates for automation.
But enterprise reporting is rarely just about putting a chart on a screen.
A real enterprise report may involve:
- Complex business rules
- Multiple data sources
- Security requirements
- Calculated measures
- Data quality issues
- Performance optimization
- Scheduling
- Bursting
- Regulatory requirements
- User-specific access
- Validation and reconciliation
AI can assist with many of these activities, but organizations still need people who understand the business and the data.
The developer’s role may therefore shift from “building every component manually” toward “designing, validating, governing, and optimizing analytics solutions.”
The Bigger Picture
Cognos Analytics 12.1.3 represents an important shift in enterprise BI.
Three capabilities illustrate this change particularly well:
Report Authoring Agent
Natural language can be used to accelerate report creation and modification.
MCP
Cognos can participate as a governed analytics tool inside broader AI workflows.
BYO-LLM
Organizations gain more flexibility in selecting the language models powering their AI capabilities.
Together, these features move Cognos beyond the traditional idea of a BI platform.
The future could look less like:
User → Dashboard → Report
and more like:
User → AI Assistant → Governed AI Workflow → Cognos → Trusted Data → Insight → Action
FAQs
1. What is Cognos Analytics 12.1.3?
Cognos Analytics 12.1.3 is an enterprise analytics platform release focused on reporting, data visualization, AI-assisted capabilities, and modern analytics workflows.
2. What is report authoring in Cognos Analytics?
Report authoring allows users to create, customize, format, and publish interactive reports using enterprise data sources.
3. What is MCP in Cognos Analytics?
MCP (Model Context Protocol) provides a standardized way for AI applications and models to interact with tools, data, and analytics capabilities.
4. What does BYO-LLM mean?
BYO-LLM means Bring Your Own Large Language Model. It allows organizations to use their preferred or approved LLM with their analytics environment.
5. What are the benefits of using an LLM with Cognos Analytics?
LLM integration can support natural-language interactions, accelerate analytics workflows, assist with report creation, and make enterprise data easier to explore.
