How to Automate Your Operations (Yalantis Story and What You Can Learn From It)

Here’s an unpopular opinion: Not every process needs to be automated.

Automation pays off only when you:

• Know which exact process is inefficient
• Understand how much it costs to run manually
• Can prove that automating it would cost less

For Yalantis, that process was resource planning. They were losing $40–50K/month due to underutilized team hours. This article describes their story to bring you a practical guide to automation that’s grounded in real numbers and results.

Yalantis in 2020: Software zoo, $40-50K/month lost, manual work

Do you know how much resource management can cost you?

Back in 2020, Yalantis did the math. They estimated a $40–50K monthly loss tied to the underutilization of people’s hours because of:

• 10+ disconnected spreadsheets for Delivery, PMs, Staffing, HR, and C-level
• 5 third-party tools (HR, time tracking, utilization) that covered ~20% of their needs
• No reliable, real-time view of who was staffed where
• Non-core activities and time-offs not reflected in resource planning
• Hours of admin work every week, pulling people away from high-impact tasks

Now, why did Yalantis start calculating all this?

Den Rudenko, Yalantis’ CEO at the time, and his team kept running into the same issues. PMs, HR, and finance weren’t productive due to lots of admin work (chasing people for updates, manually updating spreadsheets, coordinating staffing changes, creating reports). Developers were underutilized – planned for 40 hours, but working only 32 because of a day off the PM didn’t know about. Delivery leads couldn’t explain why some team members were overbooked, while others worked 30-40% of their planned time.

“The first thing where we started to crack was just too much underutilization of people's hours.”

Suspecting that resource planning was inefficient, the Yalantis team began tracking admin hours, lost time, and underutilization in, yes, another spreadsheet :) But this one was different. It was their way of testing how much their manual process cost them before trying to fix it with automation (and potentially burning money).

You could call it their manual MVP.

Yalantis in 2025: One automation software, hundreds of thousands saved

Yalantis started tracking everything manually in a spreadsheet with formulas. For each department, you should log planned hours, actual hours worked, time off, and task types per person – often across multiple projects.

A spreadsheet helped them spot underutilization before investing in automation. You can download the timesheet calculator based on this template later in this article.

Once they had hard numbers (and saw how much they were losing), they built a simple internal planner to automatically track budgets, workloads, and profit.

Two years later, that tool evolved into their own ERP system.

As the company grew, more automation needs appeared: salary reviews, time-off requests, role visibility. And for each one, the Yalantis team followed the same path:

1. Measure manually
2. Estimate the cost of inefficiency
3. Automate if the value justifies the effort

That’s how their internal planner became SHERP – a full product ecosystem for resource planning, competence evaluation, and HR management. “It covers all our needs and allows us to abandon external dependencies”, says Den.

SHERP’s three modules, HRM, Competence Evaluation, and Resource Planning, are all connected through a central data lake. This means when something changes in one module (like HR approving time off), it updates everywhere else instantly (like in staffing availability).

Within the first year of using their ERP, Yalantis’ gross profit margin grew 15–20%, and utilization jumped to 90–100%. They got a unified view of their delivery operations updated in real-time with consistent data.

Now let’s break down what they did right, and what their story teaches other IT companies about smart automation.

What does smart automation look like?

Den and his team didn’t say, “This feels inefficient, let’s buy a tool.”

They said: “This costs us $40–50K/month. Let’s figure out where it’s going wrong, what data we need, and test if automation would cost less.”

“You need to justify automation economically. Calculate how long it takes to fill out those spreadsheets for a specific process. Add up all the hidden costs of operating and delivery. If you only have 50 people and 4 projects planned until year-end, maybe you don’t need it. For us, it became critical at 120–150 people.”

Yalantis’ automation initiative paid off because they:

• Started with one painful problem → underutilized hours due to poor resource planning
• Asked: “What decisions are we currently making blindly or too slowly? What answers don’t we have?” → Who’s overloaded or idle, and why? Why are we losing margin?
• Defined what they needed to measure → e.g., billable vs non-billable hours, admin overhead, staffing gaps
• Built a simple manual MVP → a spreadsheet to track hours and utilization for a few accounts
• Calculated the cost of staying manual → $40–50K/month lost to admin work, unplanned time off, and missed revenue
• Only then automated → starting with a custom resource planner, which grew into a full internal system to improve staffing, budgeting, and profit tracking

This setup now saves Yalantis around $5K/month on admin overhead and external tools — and brings +8–12% utilization and +10–15% revenue uplift from smarter use of resources.

“I’m not sure I’d choose to build an internal tool again – it took loads of resources. But we wanted to help other companies in the same situation avoid a three-year journey (like it was for us) and automate faster where it makes sense. That’s why we turned SHERP into a full product ecosystem others can now try.”

What does it show us?

Smart (= efficient) automation starts by asking:

• What decision are we trying to improve?
• What friction are we trying to remove?
• Why does this process exist, and what’s breaking it?

For example, let’s say your delivery leads are constantly chasing PMs to check who’s available for the next project. In this case, the decision would be “Who’s available to staff next week?” Friction is time lost in Slack threads, unclear time off, and duplicated spreadsheets. The break point would be PMs planning with outdated data → people get double-booked or sit idle.

That’s why, before you automate anything, you need to slow down and spot the real issue, which brings us to the practical part: How do you know what to automate in your own company, and how do you get started?

Six areas to automate in a service-based IT company

Where are you losing time, money, or control?
What information do you need right now to run your business at scale?

Start there. That’s your signal for what might need automation.

These inefficiencies aren’t always obvious, especially when teams rely on workarounds or solve things on the fly.

But from Yalantis’ example and our own experience of working with IT service companies, here are six areas where we see automation pay off:

No matter where the problem lies, run a quick check:

• How much manual effort does this process take?
• Are we missing anything essential to run it well?
• What’s the cost of doing it wrong or too late?
• Is that cost higher than building or buying a tool?

Cases when you may not need automation

If a certain process wastes a bit of time but doesn’t cause real business damage, automating it might cost more than it saves.

Below are some examples when you can likely wait with automation:

• You’re under 50 people, managing just 3–4 long-term clients or internal teams
• Delivery cycles are stable, and team composition rarely changes
• Founders are still close to daily ops; they don’t want to step out anytime soon
• There’s little friction between HR, finance, and delivery
• Your team isn’t guessing who’s doing what

But if a process is causing consistent, measurable losses (missed hours, lost revenue, or slow decisions), automation might make sense. Just make sure it’s cheaper than sticking with the manual approach.

Don’t forget to factor in:

• Tooling costs: licenses, integrations, or building a custom solution
• Change costs: time to train the team (e.g. 2 sessions × 10 people)
• Maintenance costs: updates, in-house admin for configurations, data checks, permissions

5-step manual MVP to calculate automation ROI

As we noted earlier, before you choose to automate (if at all), you need to understand what your current process actually costs you.

Here’s how to build a simple manual MVP on the example of resource planning, inspired by Yalantis’ case.

Step 1: Choose a small sample

Pick one project and track 2–3 people’s workloads for a week. Choose a mix: one developer, one QA, one PM. This step is just to observe the process, not change anything.

Step 2: Gather your baseline data

Create a simple table in Excel to record:

• Total available hours (e.g. 40h/week, adjusted for PTO, public holidays)
• Planned leave or absences (check calendars)
• Billable hours (client work)
• Non-billable hours like internal meetings, training, and admin (if unavailable, ask for estimates)
• Project-specific hours (break down effort by task or deliverable)
• Overtime (hours beyond contract, if relevant)

Step 3: Build your MVP tracker in Excel

Structure your sheet with:

• Columns = Timeframes (days or weeks)
• Rows = Team members and task types (billable, non-billable, etc.)
• Bottom row = Formulas to calculate utilization and other key delivery metrics:

If you don’t bill clients, use productive hours instead of billable hours.

Color code your utilization levels:

• Green = underutilized (<80%)
• White = balanced (80–100%)
• Red = overbooked (>100%)

Step 4: Review the results

After a week of tracking, look for patterns. For example, who is under- or overbooked consistently? Are people spending time on unexpected internal tasks?

Example:

• QA and PM are overloaded with internal work.
• Dev A is underbooked.
• No one hits 90–100% utilization, your target for healthy margins.
• The PM spends more time coordinating people than leading delivery.

As you review, note:

• Did it take too long to gather info?
• Was someone double-booked?
• Was the data correct?

Step 5: Estimate the manual cost, effort, and risk

Now you can quantify what this inefficiency is costing you:

• Underutilization cost → Example:
  Dev A is underused by 5h/week and costs $50/hour = $250/week. Multiply by 4 people × 4 weeks = $4,000/month
• Admin overhead → Example:
  6 managers × 3 hours/week = 18 hours/week × $60/hour = $1,080/week or ~$4,300/month
• Risk of getting it wrong → Delays in project starts, misused senior talent, overworked engineers all reduce client satisfaction and margin.

Once you have these numbers, you can answer: “This process is costing us $XK/year. A SaaS tool costs $600–1,200 (or else) a month. Is it worth automating?”

Or maybe: “We’d need to build something custom to match our process. Is it still cheaper in the long run? Do we have the internal resources for this?”

[Bonus] Spreadsheet template from Yalantis’ CEO

If you’d like to shortcut this process, here’s the template built on the Google Sheet Yalantis used before creating SHERP. Just add in your company data and start tracking key metrics.

And if you need help spotting where your operations are leaking time or money, feel free to reach out. Wiseboard advisors have helped 60+ IT companies break through their growth bottlenecks, and we’d be happy to do the same for you.