Buy when the task is the same everywhere and a tool already does it well. Build when it depends on the systems, the rules or what sets the company apart. In between, the most common answer is to combine: market tools for the standard, custom work for the core of the business.
Four questions to decide
- Is the task the same in every company? Taking meeting notes, proofreading, transcribing a call: yes. Matching invoices against internal tolerances, answering according to in-house policies: no.
- Where is the data? If the agent must read and write in the ERP, the CRM and the business tools, integration will be most of the work, and a generic tool will plug in poorly.
- Which rules must it follow? Approval thresholds, exceptions, access rights: the more precise the rules, the more a standard tool forces workarounds.
- Is it a competitive advantage? What sets the company apart from its competitors deserves to be built and to belong to it. The rest can be bought.
In short
| Buy | Build | |
|---|---|---|
| Getting started | Fast | A few weeks for a first case |
| Fit with internal rules | Limited to the settings provided | Complete |
| Integration with the systems | The vendor’s connectors | Whatever is needed |
| Data | With the vendor, under its contract | Where the company decides |
| Ownership | A license | The code belongs to the company |
| Cost | Per-user or usage-based subscription | Build, then operations |
The traps
- Building what already exists. If a tool does the task well, custom work will cost more for a comparable result.
- Buying for a unique process. A generic tool on a process specific to the company often ends up worked around by the teams.
- Forgetting the run. Bought or built, an agent must be looked after: quality, costs, changes. Who does it must be planned.
Extra’s position
The diagnostic says when a market tool is enough. Only what needs building is built, then connected to the tools that stay.