By late 2021, the low-code and no-code conversation had reached a level of seriousness that earlier waves of similar ideas had not quite achieved. Microsoft Power Platform was being aggressively pushed to enterprise customers. OutSystems and Mendix had been quietly winning major customers for years. Airtable and Notion were sitting in a different but related part of the same conceptual space, blurring the line between database and document for non-technical users.
The arguments for the category were old. People with deep domain expertise who could not write code were stuck waiting for engineering teams to build things for them. If they could build small applications themselves, they could automate the work they understood best, and engineering teams could focus on harder problems.
The arguments against were also old. Software built without understanding software tends to break in ways that are hard to fix. Maintenance becomes someone’s problem eventually. Security and compliance often get treated as afterthoughts.
What had changed by 2021 was the surrounding environment rather than the tools themselves. The pandemic had accelerated digital transformation in ways that had not produced the engineering capacity to keep up. Business teams had concrete operational problems they needed solved quickly, and the budget to pay for solutions, and engineering backlogs that meant traditional development was not going to deliver in the timeframes required.
In that environment, low-code platforms found genuine traction. The successful deployments tended to share a pattern. They were used for workflows that were genuinely operational rather than core to the business. The teams using them had IT support to handle the parts that crossed organisational boundaries, like data integrations and identity management. The platforms were governed centrally so that what got built could be inventoried and supported rather than disappearing into a sprawl of unmaintained tools.
What the category had not solved by the end of 2021 was the genuine engineering work that even simple applications still required. Versioning, testing, environment management, monitoring, and the discipline around production changes were not automated by the platforms. They could be ignored, which is what often happened, with predictable consequences over time.
The interesting question for the next few years was whether the platforms would mature into real engineering tools that happened to have visual interfaces, or whether they would settle into a niche where they handled simple cases well and passed harder ones back to traditional engineering teams.