Computools and the expensive art of waiting

The printer runs out of ink precisely when someone needs the printer. This is hardly a revelation. Yet there is a business hidden inside that small irritation: a customer must notice the shortage, identify the cartridge, find a seller, pay and wait. A machine built to save effort has quietly manufactured a new errand.

Epson wanted to shorten that errand. In Computools’ account of the project, the answer was ReadyInk: monitor ink levels, notify customers and resellers, and connect replenishment with ordering and delivery. Computools supplied engineers who worked alongside Epson’s internal team. The useful invention was the connection between a warning and what happens next.

01 / THE INVISIBLE ERRANDA cartridge is a customer relationship

The Epson case makes a sprawling software company unusually easy to understand. Computools works on the distance between a business noticing something and a business doing something about it. That distance might involve a database, an interface, another company’s system, or a human decision.

For a printer manufacturer, replenishment also has commercial consequences. A customer who can obtain the right cartridge with less fuss has another reason to stay with the supplier. Resellers gain a recurring point of contact. The engineering problem and the relationship problem occupy the same desk.

Epson’s constraints were also familiar: a tight schedule, limited availability of qualified internal engineers and labor costs. Computools’ case describes an expertise-based selection of people to complement the existing team, including Java and Oracle database skills. Buying extra capacity made sense because there was already a specific job for that capacity to do.

02 / A USEFUL ADMISSIONThe founder who tried to fix everything

There is a clue to this approach in a March 2024 essay by Sergii Tymchuk, Computools’ co-founder and chief executive. Early in his business career, he says, he tried to manage many tasks simultaneously. He eventually concluded that the approach was ineffective and created further problems.

“Instead of focusing on refining a single process, I attempted to handle everything simultaneously.”

Sergii Tymchuk · management essay, March 2024

His prescription is to examine bottlenecks individually, give each proper attention and address its cause. A shortage of skilled people, for example, calls for hiring or training. Discounting the work may soothe a client while leaving the shortage untouched. He also argues for restraint: some situations need flexibility rather than another elaborate system.

The present company began in 2013 with Tymchuk, Grigoriy Shadara, Oleg Svet and Viktoriia Stepanova. Its name has an older history: Canadian engineer and businessman Roger Marion created the Computools brand in 2000. The company distinguishes that earlier brand from the business the four co-founders independently established. A name can have a past without making a young company older.

03 / SEARCH, THEN SHOPThe manual, the search box and the dress

The aviation work is a sterner version of the same problem. A technician needs the correct maintenance instructions, quickly. Computools’ Bombardier case describes persistent application bugs, an architecture needing updates and a need for better document access. Its engineers joined OBDS’ in-house development efforts; the project story is more specific than a client logo alone suggests.

The work included full-text search, indexing and APIs. Computools retained ASP.NET MVC where it fitted older applications and proposed ASP.NET Core for new cloud features. It also describes replacing an older search component with Lucene.NET. These are choices about continuity: improve what people depend on while keeping the surrounding system usable.

Now consider Dior. A shopper sees products and a checkout. Behind them sit catalogues, orders, stock movement, sales statistics and support requests. Computools’ Dior Data Hub case describes a tailored content management system to handle those operations, with a separate management panel for staff.

The reported improvement deserves its “up to.” It belongs to a particular vendor case, not a promise that every retail build will produce the same result. Still, the lesson is concrete. A luxurious shopping experience depends on rather ordinary machinery working at the right speed.

04 / THE ENGINEERING BILLYou are buying a team as well as code

Computools is a services business. A buyer can bring it an idea for a new product, an existing system needing modernization, or a delivery team missing particular skills. Consulting defines the problem and architecture. Engineering builds the solution. Dedicated teams add people to the client’s operation. Support keeps the result useful after launch.

The catalogue extends from web and mobile development to cloud migration, quality assurance, cybersecurity, data engineering and AI. Its customers include enterprises, smaller businesses and startups. The company reports more than 250 engineers and more than 400 delivered projects. Those figures describe its stated capacity and track record; they do not tell a buyer which specialists will actually be assigned.

The hourly number is only an opening question. Team size, discovery, integrations, testing and maintenance affect the total. A cheap hour spent repairing an avoidable mistake can be an expensive hour indeed. The buyer should agree what counts as completion, who owns decisions and how changes affect the estimate.

Within the market, Computools offers an alternative to building the whole capability internally or hiring another custom software provider. Its positioning combines industry knowledge with responsibility across the delivery process. Whether that distinction matters depends on the brief. If packaged software already handles the workflow well, a bespoke build needs a persuasive reason to exist.

The company’s employee-development material describes internal training and individual development plans. In February 2025, it announced that Tymchuk had joined the Employers and Entrepreneurs Council at Zaporizhzhia Polytechnic. That fits a services company’s recurring concern: the next project needs expertise, and expertise must be developed before it is urgently required.

05 / THE NEW QUEUEFaster drafts, slower decisions

Computools now describes an AI-assisted delivery model. Its argument is that analysis, planning, design drafts, implementation and testing can all move faster when teams have reusable components, clear context and defined review points. Engineers remain responsible for production work.

That qualification matters. Core architecture and security-sensitive behavior require closer human control. Isolated prototypes can tolerate more automation. Dependencies, validation and the client’s speed of decision-making influence the result. The practical question is where time disappears between the request and the accepted feature.

AI can produce a draft while everyone is still discussing who should approve it. A faster implementation exposes that delay. Computools’ delivery essay explicitly addresses approval cycles, making its current AI argument an extension of the earlier bottleneck argument.

06 / BORROW THE QUESTIONStart with the queue

A reader can use the approach before hiring anyone. Pick a recurring delay. Follow one order, document or request through the process. Record where it waits, who must act and what information is missing. Choose an outcome that a user can recognize: the cartridge arrives, the manual appears, the purchase completes.

Then test a proportionate fix. A clearer responsibility might be enough. An integration might remove duplicate work. A custom platform becomes plausible when the problem crosses systems, requires unusual business logic or exceeds the existing team’s capacity. Without a defined outcome and someone available to make decisions, adding developers may simply enlarge the meeting.

Computools’ most useful cases make the work legible in those terms. The object at the center can be spectacularly mundane: ink, a document, an order. But businesses do not lose time in the abstract. They lose it one interrupted task at a time.