Monthly AI Intelligence Report: July 2025

Executive briefing on the AI developments that matter for business decision-makers

Bottom Line Up Front: July marked the mainstream arrival of code agents, fundamentally changing how software gets built. While open-source models like Kimi K2 challenged proprietary dominance, the real transformation happened in development workflows - from advisory planning to automated code review.

---

This Month's Game-Changing AI Developments

Code Agents Go Mainstream

July 2025 will be remembered as the month code agents moved from experimental to essential. Multiple platforms launched production-ready coding assistants that integrate directly into development workflows:

• OpenAI CodeX Research Preview: Advanced code analysis capabilities, though token consumption remains expensive for extended use.

• Claude Code CLI: Game-changing integration allowing existing Claude subscribers to use AI coding assistance without additional subscriptions. Our consultant’s report significant productivity gains in code review and legacy system analysis.

• Google Gemini CLI: Beyond code generation, Gemini's CLI tool can create videos and images alongside development work, opening new possibilities for documentation and client presentations.

• Amazon Kiro: Focuses on structured coding approaches, adding rigour to AI-generated code compared to more freestyle tools.

• Atlassian Robo: After 18 months in development,it finally integrates with JIRA and Confluence, creating seamless workflow automation for teams already using the Atlassian stack

The Open Source Breakthrough

Kimi K2 Release: Chinese company Moon Shot released what many consider the most capable open-source model to date. With 1 trillion parameters (32 billion activated), it's showing performance benchmarks that rival proprietary models. Early testing shows impressive reasoning capabilities, though at 1.3TB, deployment remains challenging.

XAI Grok 4: Released ahead of schedule, possibly to counter OpenAI's rumoured open-source announcement. Focuses on reasoning and data analysis, though early reviews suggest mixed results.

Browser AI Integration

Perplexity Comet Browser: A Chrome-based browser with built-in AI capabilities for trip planning, map understanding, and spreadsheet analysis. An early signal of AI becoming infrastructure rather than an application.

---

AI Intelligence: What Our Consultants Are Actually Using

The Three-Pillar Framework in Action

Our team has evolved beyond "ChatGPT for everything" to a sophisticated three-pillar approach:

Advisory Tools (Strategic Planning):

• ChatGPT for broad market research and concept exploration

• Claude for detailed technical planning and requirements gathering

• Gemini for real-time market analysis and competitive research

Builder Tools (Development Work):

• Claude Code for legacy system analysis and code review

• GitHub Copilot for day-to-day coding assistance

• Cursor for maintenance on complex applications

• Various specialised tools (Augment, Code Rabbit) for specific use cases

  
  

Designer Tools (Client Workshops):

• Bolt and V0 for rapid prototyping with clients

• UX Pilot for wireframing and user flow design

• Lovable for deployable prototypes

Real Implementation Patterns

Multi-Tool Workflows: Rather than relying on single solutions, productive consultants chain tools together. Example workflow:

Gemini for market research → ChatGPT for strategic brainstorming → Claude for detailed technical planning → Bolt for client prototype → Claude Code for production implementation.

Client-Approved Tool Lists: Forward-thinking clients like Gold Coast Hospital now maintain approved AI tool lists, streamlining consultant onboarding and ensuring compliance.

Automated Code Review: Teams using GitHub Copilot or Code Rabbit are implementing AI-powered PR reviews as standard practice, catching issues human reviewers might miss in large changesets.

---

Industry Developments

Model Releases and Updates

• OpenAI Announcements: Despite rumours of open-source releases, OpenAI delayed due to "security considerations." Their consulting arm now publicly accepts enterprise engagements ($10-15M minimums), signalling a shift toward high-value professional services.

• Anthropic Enterprise Focus: Continuing expansion of enterprise partnerships, with Claude Code representing a significant strategic bet on developer productivity.

• Google Gemini Improvements: Extended context lengths and improved reasoning capabilities, particularly strong for real-time analysis tasks.

Regulation and Ethics

• US RAISE Act Progress: The Responsible AI Safety and Education Act continues to move through Congress, establishing mandatory transparency standards for AI systems.

• Getty Images vs Stability AI: Ongoing lawsuit regarding image training data, with implications for all generative AI models using copyrighted content.

• EU AI Act Enforcement: Unexpected delays in enforcement due to industry lobbying, creating uncertainty for organisations operating in European markets.

Client Action Items for August

Immediate (Next 30 Days)

✓ Evaluate current development tool licensing: If you have GitHub Copilot, test automated PR review capabilities

✓ Assess code agent readiness: Identify legacy systems that could benefit from AI-assisted analysis

✓ Review AI tool approval processes: Establish clear guidelines for consultant tool usage

Strategic (Next 90 Days)

✓ Pilot the three-pillar framework: Test advisory, builder, and designer tools for specific project phases

✓ Develop AI-enhanced client workshop processes: Integrate real-time prototyping tools into discovery sessions

✓ Plan for code agent integration: Budget for training and tool licensing as code agents become standard practice

Long-term (Next 6 Months)

✓ Establish AI governance frameworks: Create policies addressing tool selection, data handling, and quality assurance

✓ Invest in team AI literacy: Ensure technical and non-technical staff understand appropriate tool selection

✓ Prepare for workflow transformation: Plan organisational changes as AI tools reshape development processes

---

August 2025 Predictions

Code Agent Standardisation: Expect major IDEs to announce built-in code agent capabilities, making AI-assisted development the default rather than optional.

Open Source Model Proliferation: Following Kimi K2's success, anticipate more capable open-source models challenging proprietary alternatives.

Browser AI Integration: Perplexity Comet represents an early signal of AI becoming browser infrastructure; expect major browser vendors to follow.

Enterprise AI Consulting Boom: As organisations struggle with tool selection and implementation, demand for AI strategy consulting will surge.

The Reality Check

Despite impressive technical advances, successful AI implementation remains grounded in business fundamentals:

What's Working: Organisations treating AI tools as workflow enhancers rather than replacements, maintaining human oversight in critical decisions, and focusing on measurable productivity improvements.

What's Not: Companies chasing every new tool without a strategic purpose, implementing AI without governance frameworks, and expecting tools to solve process problems.

The most successful implementations enhance existing workflows while maintaining quality standards and human judgment.

---

What AI developments are you tracking for your organisation? How are you approaching code agent evaluation and implementation?

Want monthly AI intelligence from software consultants implementing these technologies daily? Follow our insights for practical guidance from the field.