Enterprises face the need to decide: build their own systems, acquire ready-made solutions, or use a hybrid approach combining both options. Ready-made platforms offer proven functionality and speed of implementation, while custom-developed solutions allow for a higher level of control and customization. Increasingly, large companies are opting for a hybrid strategy, collaborating with specialized partners to create scalable systems.
Matthew Barnard, CEO of BBD Software, notes that clients are in a phase of active experimentation and learning. He emphasizes that the pace of change is so high that knowledge gained two weeks ago might be obsolete in two weeks, and the company cannot remain detached from this process.
There is no universal answer to this question; the right choice depends on numerous factors, including context, regulatory requirements, and the technology's potential in creating a real competitive advantage.
Thanks to artificial intelligence, the barriers to software development have significantly decreased. Organizations can now adapt open-source models and apply AI assistance in development for accelerated prototyping and iterations. The 'vibe coding' technology, where developers use natural language prompts for AI tools to generate and refine applications, has made projects previously considered too expensive or time-consuming feasible.
In the financial services sector, this even changes the economics of modernizing legacy systems. Barnard cites examples of Cobol systems still widely used in banks. Projects requiring years of overhaul are becoming more realistic. According to him, some estimates provided to clients are half the cost they would have been three years ago.
Theodora Lau, founder of Unconventional Ventures and author of Banking on (Artificial) Intelligence, argues that this shift will force software providers to reconsider pricing, as organizations often only require part of the SaaS product's functionality.
However, faster and cheaper creation does not eliminate the complexities associated with deploying enterprise software into a working environment. Custom developments still require management, maintenance, and security. In the financial sector, accuracy is an absolute prerequisite.
Barnard points out that SaaS platforms contain accumulated knowledge and intellectual property that cannot simply be recreated through faster coding. Lau agrees, stressing that when dealing with people's money and regulators, a 10% error tolerance is unacceptable—100% accuracy is required.
The strongest argument for customization may lie in areas where organizations genuinely seek to stand out. Barnard and Lau believe that highly regulated core systems, such as payments and accounting, will remain standardized. Client interfaces offer more opportunities for innovation.
Barnard advises enterprises to determine where they see a technological advantage. He adds that companies look for ways to do something unique that competitors have not done, and this most often happens in customer-facing interactions, rather than in backend accounting systems based on international standards.
In Barnard's view, the first 90% of the solution can be built quickly, but the main work is concentrated in the last 10%: operation in the working environment, scaling, support, and maintenance. Lau shares a similar thesis regarding the need for human oversight. Although AI can automate processes related to documents or assist in credit decision-making, organizations must be able to trace the decision-making process.
The next major shift is towards agentic systems. As AI integrates into development and decision-making, organizations will have to move from sporadic experiments to ensuring governability, repeatability, and accountability. Regulators are already paying more attention to this. The EU AI Act imposes strict obligations on high-risk systems, while the US and UK adopt different approaches.
Lau expects increased requirements for audit trails, explainable results, and evidence of risk reduction as agentic commerce grows. Furthermore, firms will need to know which specific agent initiated a transaction, what authority it was granted, and how it is verified. Barnard notes that due to the speed of change, long-term certainty is difficult. He believes it is better to move quarter by quarter, constantly learning and evolving. BBD is experimenting with specialized AI agents capable of replicating elements of the work of developers, business analysts, and project managers. Barnard concludes that these agents are orchestrated to perform various tasks, and this is done very cautiously, as only high accuracy is acceptable at the enterprise level.
A major announcement has been made regarding the Namo Bharat RRTS corridor between Ghaziabad and Jewar. NCRTC has unveiled the names of the stations that will be constructed along this 72.44-kilometer corridor.
Proposed stations include Ghaziabad, Greater Noida, and Noida International Airport. NCRTC specified that there will be 11 stations on the Namo Bharat route, whereas 18 stations are planned for the metro to be built on the same stretch.
This infrastructure will allow passengers to travel between Ghaziabad, Noida, and Jewar Airport both by regional high-speed train and local metro.
NCRTC, or the National Capital Region Transport Corporation, is a government entity responsible for developing and operating RRTS corridors in Delhi-NCR. However, one official stated that this corridor is currently in the detailed project report (DPR) stage.
The 11 stations listed for Namo Bharat on this route are: Ghaziabad South, Greater Noida Sector 2, Greater Noida Sector 3, Greater Noida Sector 12, Surajpur, Alpha-I, Ecotech-VI, Dankaur, YEIDA North (Sector 18), YEIDA Central (Sector 21), and Noida International Airport.
The names of the 18 stations planned for construction on this route are also presented: Siddharth Vihar, Greater Noida Sector 16C, Ecotech 12, Greater Noida West (Sector 4), Greater Noida Sector 10, Gyan Park V, Police Lines Surajpur, Ecotech-2, Gyan Park III, Omega-II, Phase-III, Ecotech-1E, Ecotech VIII, Dadupur, Junedpur, YEDA Sector 17, YEDA Sector 15, and YEDA Sector 33.
It is planned that trains on this corridor will travel at a speed of 180 km/h, which is expected to reduce the travel time between Ghaziabad, Noida, and Noida International Airport to approximately 40-50 minutes.
The official noted that the goal of this project is to strengthen the transport accessibility of the entire National Capital Region by connecting Ghaziabad with Greater Noida, the Yamuna Expressway area, and the future Noida Airport. Upon completion, this corridor will provide an additional public transport option for passengers traveling between Ghaziabad, Noida, and Noida International Airport, improving connectivity with numerous residential, commercial, and developing infrastructural centers along the way.
A working group, including representatives from the management of the Main Directorate for Transport Development and Road Transport Infrastructure of Tashkent, the interdepartmental operational headquarters, and JSC 'Toshshahartranskhizmat', conducted an inspection of the capital's transport infrastructure facilities.
As part of the inspection, specialists closely examined bus stops located on Fargona Yuli Street to assess the level of landscaping and passenger conditions. Subsequently, the commission proceeded to the 1st Bus Park, where the technical condition of the rolling stock and the maintenance of the buses were evaluated. Instructions were issued to the responsible employees regarding the urgent elimination of the identified shortcomings.
During the inspection, the final stop named 'Aviasozlar' was also checked. The tour concluded on Shota Rustaveli Street, where construction work is underway to expand the dedicated lane for BRT. Specialists inspected the ongoing projects and determined tasks aimed at ensuring the safe and convenient operation of public transport.
CodeRabbit has successfully raised $143 million in a Series C funding round, valuing the company at $1.5 billion. Atomico and Smash Capital participated as co-leads in this round. New investors include BMW i Ventures, Datadog, Hirtle Callaghan, and SineWave Ventures. Existing investors, such as CRV, Scale Venture Partners, and Flex Capital, also supported the deal.
This funding was secured less than a year after CodeRabbit raised $60 million. The company reports that its revenue has increased more than fivefold compared to the previous year. Currently, CodeRabbit performs over two million code checks weekly for 17,000 clients, including companies like Adyen, BMW, Indeed, JFrog, NVIDIA, and Trivago.
The capital raised will be used to accelerate the company's international expansion and increase research and product development expenses. CodeRabbit plans to invest over $10 million to support open-source projects, which will allow it to maintain free access to AI review and agent features for accompanying projects.
AI coding tools enable engineering teams to create software at unprecedented speeds, as developers can generate code, create pull requests, and troubleshoot much faster. Nevertheless, every change introduced still requires testing, verification, and proper management before going into operation.
CodeRabbit asserts that this shift is transforming software development management practices in organizations. Traditional task trackers were designed to work with planned work and developer-driven execution. However, AI-based development allows code to appear before teams achieve full consensus.
This makes the pull request an increasingly critical control point. CodeRabbit CEO, Hardjot Gill, noted: 'Writing code has become dramatically easier. Verifying its quality remains complex.' He emphasized that changes created by AI still require independent verification before being released into the production environment.
CodeRabbit's goal is to provide this level of control for software created by both humans and agents. The platform analyzes changes using repository context, organizational standards, and testing evidence. The system also indicates the scope of impact and architectural dependencies that require further attention.
Along with the latest funding announcement, CodeRabbit introduced Agentic Change Management. This system expands AI review capabilities, including prioritization, explanation, and continuous code security monitoring.
Its purpose is to help engineering teams understand exactly which changes require attention. The CodeRabbit Triage tool ranks incoming pull requests based on value, urgency, and risk. It also considers reviewer dependencies, readiness, and suitability before routing the task. Lower-risk changes can pass through automated workflows, while complex changes are directed to human reviewers. CodeRabbit Change Stack provides deeper explanations of software changes and their potential impact by analyzing contracts, domain behavior, integrations, tests, and migrations between repositories.
CodeRabbit is also expanding its presence in security with an AI-powered product. This system continuously scans production code for vulnerabilities and emerging risks. It verifies detected issues against code evidence before recommending a fix via pull requests. This approach targets vulnerabilities that traditional rule-based tools might miss, including authorization bypasses, insecure object references, and business logic errors. The company believes that continuous monitoring will become increasingly significant as AI generates more software.
The latest funding will also support CodeRabbit's international expansion strategy. The company recently opened an office in London to serve corporate clients in Europe and plans further expansion in Europe, Japan, and other Asian markets. Luca Eisensteken, a partner at Atomico, will join the CodeRabbit board of directors. He stated that the infrastructure for independent governance will become increasingly important for AI-generated software. CodeRabbit believes that its model-agnostic approach will help organizations maintain software quality as AI adoption accelerates.
