The neatest cloud migration is sometimes the one that leaves things behind. In its 2021 account of work for financial technology provider Kurtosys, Cloudreach described an assessment of 58 applications. The proposed destination for 39 was the cloud. For the other 19, it was retirement. Someone had asked the useful question before the expensive work began: what actually needed to survive the move?
That is a good introduction to Cloudreach, the London-founded consultancy later acquired by Atos and brought into Eviden. It works in the gap between an enterprise deciding to use public cloud and its employees being able to run a dependable business on it. The inventory, permissions, application dependencies and operating responsibilities all need attention. A cloud account alone settles none of those questions.
Thirty-nine to move. Nineteen to lose.
Kurtosys already had a capable technology department. It wanted specialist scrutiny of its AWS architecture and economics. Cloudreach split the engagement into governance and an application business case, using Cloudamize to help inventory the estate. Its six-week governance review covered nine domains, including identity, networking, security, automation and recovery. The assessment distinguished migration effort from ongoing running costs.
The case study says Kurtosys accepted the recommendations and deployed its platform in a new Frankfurt production environment within a month of the engagement. The practical lesson is straightforward: document what exists, identify what can go, then price the remaining work. That sequence makes an assessment useful to the finance team as well as the engineers.
A company built around somebody else’s computers
James Monico and Pontus Noren founded Cloudreach in 2009. By 2015, an AWS interview with Monico described the move from a two-person operation to a business employing 150 people. His fellow founder’s interview emphasized international reach and varied customer use cases. These were the early contours of the business: learn the platforms deeply, then help enterprises apply them.
Cloudreach’s customers have included bp, Hearst, Pearson and Volkswagen Financial Services. In Blackstone’s 2017 investment announcement, its stated customer focus was mainly enterprises with more than $1 billion in revenue. Those are organizations where replacing infrastructure also means coordinating departments, legacy applications and decisions that have accumulated over years. A technical answer has to survive an organizational meeting.
Its position was that of an intermediary with engineering responsibility. The cloud providers supplied the underlying platforms; Cloudreach helped customers adopt and operate them. It could compete for work against specialist consultancies, large systems integrators or an internal platform team. The meaningful comparison is what each can assess, deliver and support for the particular estate, with named people accountable for the handover.
The software inside the consultancy
Blackstone acquired a majority stake in February 2017. The stated ambitions included geographic expansion, software development and more automation in managed services. In August, Cloudreach acquired a majority stake in Philadelphia-based Cloudamize. That deal gave a specific form to the software strategy: automate discovery and analysis that would otherwise absorb consulting hours.
Cloudamize’s job was to help identify applications and infrastructure, map migration plans and compare cloud configurations. In the announcement, Noren connected the acquisition to the priority of putting software behind Cloudreach’s offerings. It is an important distinction in a services business. Repeatable analysis can help a team approach the next customer with a working tool rather than an empty spreadsheet.
The following year, Cloudreach announced its proposed acquisition of Atlanta-based Relus Cloud. The stated attraction combined North American delivery capacity with data analytics and machine learning expertise. Its announcement explicitly addressed the space between smaller consultancies and traditional global integrators: customers wanted specialist knowledge with enough people to deliver large projects. The deal was intended to widen both the skill set and the geographic reach.
At bp, the groundwork for AI
Microsoft’s account of Cloudreach’s work with bp makes the engineering role tangible. Cloudreach deployed and configured Azure services in bp’s upstream digital factory. Within six weeks, the company was ready to implement Azure AI and machine learning capabilities. The work included automation patterns and reusable templates, giving subsequent projects foundations they could use again.
“The spectrum of business challenges is enormous.”Lance Goodship · Principal Architect for AI, bp
The published examples require careful reading. Moving one desktop solution to Azure reportedly produced more than $2 million in savings. A document-extraction accelerator had potential value above $20 million. Another testing pattern had the potential to shorten well analysis from three months to less than a week. The first is reported savings; the others describe opportunities. They should not be added together as money already earned.
For a buyer, the interesting deliverable is the reusable pattern. Once a governed route to deployment exists, another team can start further along. That is the appeal of cloud engineering in this setting: create a foundation that makes the next useful experiment easier to conduct, test and operate.
The monitor that missed too much
A Splunk customer story supplies something more revealing than another partner badge: an account of a tool that no longer covered Cloudreach’s needs. Its previous monitoring covered infrastructure but missed application and business metrics. Sampling intervals were too long for an accurate performance picture, alerts had limited tuning and traces were absent. The view of customer environments was fragmented.
Following a competitive review, Cloudreach standardized on SignalFx infrastructure monitoring and microservices application performance monitoring. The reported benefit was a combined view of infrastructure, applications and business measures, with faster visualization and troubleshooting. This was a change in what the operations team could observe, not simply a replacement dashboard.
The lesson travels well. Before replacing an operations tool, write down what it cannot show you. Then test whether the replacement closes those gaps. A system can appear healthy at the server level while the application is disappointing the people using it. The monitoring decision needs to start with the service people experience.
A bill needs a name attached
Financial visibility was another recurring problem. In its CloudHealth case study, Cloudreach described customers struggling to attribute resources to projects and connect spending with performance. A cheaper configuration is not necessarily a useful optimization if it harms the application. Billing information and operational information need to be considered together.
CloudHealth helped Cloudreach group costs by business unit and act on resources with missing or incorrect tags. Some customers could see spending at the individual developer level. That makes accountability concrete: a resource belongs to a team, and that team can decide whether its cost serves a purpose.
A separate Cloudreach case study reported a 36 percent reduction in annual data-center costs for an unnamed US mortgage lender after an Azure migration met its exit deadline. It is a customer-specific outcome, with a defined cost category. It should prompt a buyer to ask which costs their own project will remove, and when those expenses actually stop.
- 01 / DiscoverInventory applications and dependencies.
- 02 / DecideRetire, move or modernize each workload.
- 03 / AccountSeparate transition costs from running costs.
- 04 / OperateAssign owners; watch cost and performance.
The contract is part of the architecture
Cloudreach’s commercial model combines professional services work with recurring managed operations and software-enabled delivery. The historical Infrastructure Reliability specification refers to monthly charges and an order form that determines scope. Buying the service therefore involves defining the environment and the responsibilities, as well as agreeing the commercial terms.
One detail deserves attention: the specification distinguishes environment-level support from application administration. Keeping a managed database service configured correctly does not mean administering everything inside the database. That boundary matters when a customer expects one team to solve every problem bearing the word “cloud.”
The same document places conditions on continuous improvement, including a stated monthly-spend threshold of $25,000 unless otherwise agreed, and limits on included work. That threshold is an eligibility condition, not the consultancy’s price. The buyer’s copyable move is to put infrastructure, application, security and recovery ownership in writing before signing. If the required work falls outside the support scope, the operating arrangement needs to change.
Cloudy lunches, serious recruiting
Cloudreach’s early workplace had a playful vocabulary. Best Companies described “Cloudy lunches” at local restaurants and quarterly Skunkworx days for creative projects. It also reported a deliberate avoidance of senior and junior job titles, reflecting the founders’ concern that people could end up pursuing labels for their own sake.
“People come to work for a company, not for a salary only.”Pontus Noren · Co-founder, quoted by Best Companies
These are accounts of the founder-era company, rather than promises about employment today. Still, they explain an ingredient in a consultancy’s growth: the people doing the work need reasons to stay. Shared meals are easy to imitate. Giving engineers time to explore an idea, and making advancement credible without elaborate titles, asks more of management.
From specialist to part of Atos
Atos announced its agreement to buy Cloudreach in December 2021, describing expected calendar-year revenue of approximately €100 million. It cited more than 1,000 certifications across AWS, Azure and Google Cloud, along with Cloudreach’s 2021 AWS Consulting Partner of the Year award for the UK and Ireland. Those figures belong to that acquisition-era snapshot.
The transaction closed on January 3, 2022, with completion announced the following day. Atos subsequently reported 742 employees joining through Cloudreach in its first-quarter results. In April 2023, Cloudreach became part of Eviden. Its official YouTube channel records that transition; by September 2026, the Eviden legacy-brand page redirected to Atos’s website.
The name has travelled through a changing corporate structure. The customer work remains the clearest explanation of what Cloudreach contributed: judgment about what to move, engineering to make it usable, and operations to keep it accountable. Before commissioning the next migration, a buyer could do worse than return to those 19 applications at Kurtosys. Ask what can be retired before asking how fast everything can move.