One of Sphere's projects is a CRM system for internal collaboration between employees of an agro holding. This isn't CRM in the usual "sales pipeline and customer records" sense. It's an internal enterprise system whose users are employees across the holding's different divisions who need to work together.
Using this project as an example, we look at when a company needs custom CRM development and when an off-the-shelf product will do, what makes agribusiness processes different, and how we approach projects like this, from discovery to rollout and staff training.
Custom CRM or off-the-shelf: where the decision starts
The first question to ask, before any talk of development, is whether you need it at all. There are plenty of ready-made CRMs and corporate portals on the market, such as HubSpot, Pipedrive, or Salesforce, and for many companies they're the right choice.
| Off-the-shelf CRM | Custom development | |
|---|---|---|
| Processes | Standard: deals, tasks, contacts, documents | Industry-specific, don't fit templates without workarounds |
| Company structure | Few employees, simple structure | Many divisions and sites with different roles and permissions |
| Integrations | With popular services via ready-made modules | With several internal, proprietary, or industry systems |
| Customization | The company adapts its processes to the product's logic | Needed when customizing the box costs more than the product and updates break the changes |
| Data and roadmap | Follows the product's logic and plans | The company controls where data is stored and how the system evolves |
| Getting started | Faster | Slower, but the system is built around real processes |
Giving an honest answer to this question is part of our job. Sometimes discovery shows the problem is better solved by configuring an existing product, and then it's worth saying so right away.
Why it's hard for an agro holding to find a ready-made solution
A CRM for an agro holding runs into specifics a typical office-based company doesn't have.
- A distributed structure. Head office, production sites, fields, warehouses, logistics: divisions are in different places and work at different rhythms.
- Diverse users. One system is used by executives, accounting, agronomists, mechanics, procurement staff, and drivers. They have different levels of digital skill and different devices, from a work computer to a smartphone in the field.
- Seasonality. During planting and harvest, the load on staff and on the system spikes, and there's no time to figure out a complex interface.
- Connectivity. Internet at remote sites can be unreliable, and that needs to be factored into the design.
- Lots of approvals. Purchasing, equipment repairs, and moving resources between divisions often require a chain of approvals involving several people.
Off-the-shelf products are usually built for an "average" company. The further your processes are from that average, the more configuration, customization, and compromise they require.
Off-the-shelf products are built for an "average" company: the further your processes are from that point, the more customization and compromise you need.
What an internal system for employees usually needs to do
The feature set is defined after the process review, but internal systems of this kind have some standard building blocks:
- Tasks and assignments with owners, deadlines, and statuses, visible to everyone involved in the process.
- Requests and approvals: approval routes that mirror the company's real logic rather than imposing someone else's.
- Directories and company structure: divisions, employees, roles, contacts.
- Roles and access rights: everyone sees what they need for their work, and nothing extra.
- Notifications in a convenient channel, so nobody has to keep the system open all day.
- Documents and history: attachments, comments, a change log, so that six months later you can see who made a decision and why.
- Reports for managers: where processes are stalling, which divisions are overloaded.
In an internal portal, all these blocks serve one goal: less back-and-forth in messaging apps and fewer "where's my request?" calls, more transparency.
Discovery: mapping processes, not a feature list
Our first phase is discovery, structure, and meaning. For an internal system, that means a review of the company's business processes.
During discovery we work through:
- which processes the system should support first, and where the most time is being lost today;
- who takes part in each process, what decisions they make, and what information they need to make them;
- which systems are already in use (ERP or accounting, HR, email, messaging) and which ones need to exchange data;
- what data needs to be migrated from old spreadsheets and systems;
- how success will be measured: approval speed, the share of processes run in the system, employee satisfaction.
The outcome of this phase is a description of processes, roles, and requirements, which becomes the basis for the technical specification and the work plan. It's convenient to split the project into phases: launch the most painful processes first, then expand the system.
Design: a system people want to work in
In projects like this, it's important to:
- design each role's home screen so people immediately see their own tasks and requests;
- cut down the number of steps for frequent actions: creating a request, approving it, marking it done;
- think through the mobile scenario for people who don't work at a desk;
- keep a single visual language, so new modules don't turn the system into a collection of disconnected screens.
We work through scenarios on prototypes and test them with future users before development. Fixing a prototype is cheaper than reworking a finished module.
Development, integrations, and testing
Custom CRM development usually runs in iterations: each phase is brought to a working state and tested with real users.
Architecture built for growth
An internal system almost always grows: new processes, divisions, and sites appear. So in projects like this it's important from the start to design an architecture where new modules plug in without reworking existing ones. That applies to the data model, the roles, and the approval routes, which are best made configurable so that a change in policy doesn't trigger a new development cycle every time.
Integrations
Much of an internal system's value depends on how well it's connected to the rest of the company's IT landscape. Syncing employee and division directories, exchanging data with the ERP or accounting system, single sign-on: all of this removes double entry and errors. The set of integrations is defined during discovery.
Data migration
Old requests, directories, and documents need to be migrated carefully, so employees don't start from a blank page and don't keep old spreadsheets running in parallel.
Testing
Beyond checking features, it's important to test access rights: each role should see exactly what it's supposed to. Load scenarios are tested separately, since user activity rises noticeably in peak seasons.
Rollout and training: where the project succeeds or fails
An enterprise system can be technically flawless and still never get adopted. Employees are used to their own tools, and switching takes effort. That's why, for internal systems, the launch and growth phase matters as much as development.
An enterprise system can be technically flawless and still never get adopted.
- Pilot. Launch with one division or process, collect feedback, and make improvements before scaling.
- Training. Short guides for each role, training sessions, and go-to people inside each division who can answer questions.
- Rules. Management agrees that the process lives in the system, not in a messaging app. Without that, adoption stalls.
- Growth. New processes, modules, and integrations as the system becomes a real working tool.
Staff training and post-launch development are part of how we approach web services as a digital agency: we see launch as the start of a product's life. Over time, an internal system can be extended with smart search or an assistant that answers questions about company policies.
You can find more of our projects in our portfolio.
FAQ
What makes a custom CRM better than an off-the-shelf one?
A custom system is built around the company's real processes, not the other way around. It isn't limited by a ready-made product's logic, it's easier to integrate with internal systems, and the company decides how it evolves. But an off-the-shelf product is faster to start with, so the choice depends on how far your processes are from the standard.
Can we start with an off-the-shelf CRM and move to a custom one later?
Yes, that's a common path. Working with an off-the-shelf product helps you understand which features are missing and which processes really matter. When you switch, the key is to plan data migration and staff training in advance.
How long does custom CRM development take?
It depends on the number of processes, roles, and integrations. It usually makes sense to split the project into phases: the first covers the most painful processes and launches sooner, and the next ones are added as adoption progresses. A precise plan comes after the discovery phase.
Does this approach work for companies outside agribusiness?
Yes. Distributed structures, many roles, approval chains, and seasonal peaks are common in manufacturing, logistics, construction, and retail. The logic for choosing between off-the-shelf and custom, and the order of work (discovery, prototypes, phased development, a pilot, and training), are the same for them.
How do you get employees to actually use the new system?
Involve future users from the design stage, keep the interface simple for every role, launch through a pilot, train people, and lock in the rule that the process is run only in the system. It's important that managers use it themselves.




