PALO IT wants to fix the work around the code

A website feature at Singapore Airlines reached user acceptance testing in five weeks against a nine-week forecast. The work involved one digital product squad, a familiar coding tool, GitHub Copilot, and an unfamiliar set of working habits supplied by PALO IT. For a consultancy selling artificial intelligence, it is a useful opening exhibit: a defined piece of software, an expected timetable and a result someone can interrogate.

The comparison comes from PALO IT’s account of the pilot. It is a forecast-versus-delivery comparison, not a controlled experiment. Still, it puts the conversation somewhere useful. Before buying the future of work, a customer can ask which work actually changed.

PALO IT designs digital products, builds software, modernizes existing systems and helps organizations adopt AI. Its customers include airlines, insurers, retailers and financial institutions. The pitch increasingly centers on Gen-e2, short for Generative AI Enhanced Engineering: a method for changing how a product team delivers, from its initial brief through engineering and validation.

The assistant arrived. The bottleneck stayed.

The company’s own origin story for Gen-e2 begins with disappointment. In a February 2025 account, CTO Dimitri Baikrich described introducing AI coding assistants into a conventional Agile development process. Individual developers gained help, but the broader results were underwhelming. Faster activity inside separate teams left the surrounding production chain largely intact.

PALO IT’s response was to work on shared project context: business goals, architecture choices, implementation decisions and reusable instructions. Its account describes 18 months of experimentation. Product owners, designers and engineers began working together in AI-assisted group sessions, with the system carrying a fuller record of what they were trying to build.

The distinction matters commercially. A buyer can purchase a coding assistant directly. Hiring PALO IT means buying help with the organization around it: deciding what to build, connecting the tools and changing how specialists work together. The consultancy must therefore demonstrate value beyond access to the tool itself.

A shorter runway, with a clearly marked finish line

At Singapore Airlines, the initial target was a website feature ready for acceptance testing. PALO IT also reports improved code coverage and a build success rate above 90 percent three weeks into the pilot. Positive results led to plans for broader testing and adoption.

That finish line deserves to stay visible. Acceptance testing is a milestone before a feature’s ultimate business effects can be judged. The five-week result says something about delivery; it does not establish a gain in passenger satisfaction or revenue. Keeping those questions separate makes the engineering story more persuasive.

Australia supplies a different example. HCF selected Active Locals, a mobile app designed to help people find activity groups suited to their interests and fitness levels. PALO IT says it delivered the app in four months with half the engineering team anticipated under a traditional Agile approach, while adding features beyond the original scope.

The case study puts the productivity gain at 86 percent. That is the project’s reported comparison, not a standard return attached to every engagement. The more tangible result is an app for finding people to get active with, built by an insurer exploring how social connection could support healthier lives.

The gym app’s heavy lifting happened underneath

Virgin Active Australia needed to align membership systems and support a mobile app. The critical work sat in middleware called va-api, the connecting layer between systems. PALO IT modernized that codebase and adapted it to local business requirements while supporting integration with the global app.

The company reports a successful membership migration and 40 percent cost savings by the project’s end. It also says features such as crèche bookings and personalized training programs could be introduced in hours where they previously took weeks. For a club member, the reward is a booking that works. The engineering effort stays backstage.

For buyers, a percentage saving is the start of a procurement conversation. They should define the comparison budget, the scope included and the cost of keeping the system running. A project can save engineering effort while still requiring integration, training and careful review. Those expenses belong in the same calculation.

A consultancy with its own machinery

PALO IT’s business combines consulting, custom delivery and training, alongside its Impact Tracker software offering. A customer might start with product discovery, commission an application, modernize a platform or train an existing workforce. That mix places it in the enterprise engineering market, where an internal team is also a serious alternative.

Gen-e2 gives that services business a named method and reusable assets. Its current Harness offering includes setup packages, a shared product workspace, an enterprise AI hub and agents configured for tasks such as requirements, testing and review. Human approval and validation sit within the proposed workflow. The method’s ambition is to make delivery practices repeatable across projects.

Its relationship with GitHub is unusually visible. GitHub named PALO IT its 2025 APAC Services and Channel Partner of the Year. The same awards recognized Accenture and Avanade, Capgemini and Cognizant in other categories. That is a useful map of the neighborhood: established services businesses competing to help enterprises put a common technology platform to work.

PALO IT’s position rests on the combination of engineering practice, organizational training and sustainability work. None of those activities alone is exclusive. The test is whether combining them helps a particular customer complete a difficult project with less friction.

Buying a school for the AI era

In March 2026, PALO IT announced its acquisition of Bangcast, an AI training specialist whose reported track record included more than 6,000 professionals across roughly 100 organizations. It brought the business into PALO IT Academy, expanding the consultancy’s capacity to teach the people expected to use the systems.

“Deploying AI without redesigning the organization is the fastest path to disappointment.”

Stanislas Bocquet
Founder and CEO, on the Bangcast acquisition

Colissimo’s training case makes that concern concrete. The parcel business had employees with different levels of AI experience, strict data requirements and restrictions on mainstream tools on company computers. The program began with interviews, surveys and actual working documents. Six-hour group sessions used tasks such as meeting preparation, translation and spreadsheet analysis.

The published results cover more than 300 employees and report average savings of three hours a week per participant. These are training outcomes reported in the case study. The detail worth copying is the preparation: start with a real document and a recurring job, then teach the skill in that setting.

At JLL, the Academy trained 105 employees using exercises tied to real estate work and internal tools. Participants left modules with two or three use cases to apply. At Pernod Ricard Mexico, a separate program trained 266 employees and produced more than 600 specialized prompts and 240 identified implementation opportunities. An opportunity is still a proposal; the distinction keeps a training tally from masquerading as completed automation.

In June 2026, PALO IT announced Microsoft Training Services Partner designation. The Academy’s offering now joins instruction, practical skills and certification pathways. That broadens the company’s relationship with a customer beyond the team writing software.

The green work has a data problem to solve

PALO IT’s sustainability business has its own concrete product. Impact Tracker brings environmental, social and governance information into a platform with permissions, audit logs, calculation transparency and data integrations. The practical aim is to make information easier to trace, compare and use across an organization.

Its carbon-market work is different again. PALO IT co-designed and built CIX Exchange, a spot-trading platform for Climate Impact X, serving participants such as financial institutions and commodities traders. The work concerned a usable, extensible trading system. Building market infrastructure and delivering an environmental outcome are separate tasks, even when the first is intended to support the second.

The company’s B Corp identity has an independently checkable basis. B Lab lists its Singapore operation as certified since April 2019, with an overall impact score of 95.1. Certification belongs to particular entities; it should not be treated as a stamp on every software project. PALO IT’s own B Corp page describes training, governance practices and greenhouse-gas measurement as part of its operating commitments.

What survives the next change in tools?

Stanislas Bocquet founded PALO IT in Paris in 2009 with a small team of senior architects and technical leaders. Singapore followed in 2012, Hong Kong in 2014 and Australia in 2015. The company now lists Hong Kong as its headquarters and works across APAC, Europe and the Americas. Its name borrows from Palo Alto’s innovation culture; IT also carries the meaning Innovation & Transformation.

The useful lesson across its recent work is a bounded experiment. Choose a product problem, record the requirements and agree on a baseline. Include the people who approve the work. Count review effort and defects alongside speed. These are editorial takeaways from the projects, not a promise that another team will reproduce their percentages.

The approach depends on people being able to change their workflow, use approved tools and validate the output. Without those conditions, the earlier disappointment is easy to imagine returning: a faster task inside a project that still waits. PALO IT’s wager is that customers will pay to shorten that wait. The next useful proof will be another finished piece of work.