91制片厂 has reached a pivotal moment, driven by rapid changes in AI and the needs of privacy-first organizations. To continue supporting these teams, we are evolving. Here is our approach.
In my last post, I compared Lowcode to AI Code generators. Now, I鈥檒l focus on 91制片厂 - what we believe, what we鈥檙e doubling down on, and where we鈥檙e heading.
What We Stand For
These are the principles that guide how we build 91制片厂:
- Privacy. Keep your data within your own infrastructure and trusted providers.
- Ownership. Stay in control of your data and the software that processes it.
- Connectivity. No walled gardens - connect to the tools that you trust.
- Humans in the loop. The future is AI Assisted. People still make the critical decisions.
91制片厂 has always been built for organizations that prioritize security, compliance, and data residency. As AI becomes foundational to modern workflows, our mission is simple:
Enable privacy-first teams to adopt AI with confidence
AI Workflows
Today, you can build low-code apps and automations with 91制片厂. That won鈥檛 change. In fact, automations and apps will become even more powerful.
We鈥檙e betting on a future where AI assistants will be embedded in most workflows. To support this, we鈥檙e adding two huge pillars to 91制片厂 Workspaces:
1. Agents
AI Agents work best when they are single-purpose - where each agent has a clear set of instructions and a well defined set of tools to perform its tasks.
Inside 91制片厂, you鈥檒l be able to create multiple specialized agents - each with access only to the data and APIs you choose. The chat interface (or another agent) will help orchestrate which agent to call for each request.
Agents will be usable via automations and the new chat interface (below).
And, most importantly, Role-based access controls will apply.
- No agent can bypass your permissions.
- No human can see or do anything they have not been given access to.
91制片厂 provides the foundation for secure AI automation in regulated environments.
Models
91制片厂 will support a range of LLMs via , so you鈥檒l be able to use any model that鈥檚 compatible with the OpenAI format. For beta, our testing will focus on smaller, open-source models:
- Mistral: ministral-14b-2512
- Qwen: qwen3-32b
- OpenAI: gpt-oss-20b
- Meta: llama-4-scout
91制片厂 agents will also function well with the latest models on all the major public providers, including OpenAI, Gemini, Mistral, Anthropic, etc.
2. AI Chat
You already know how to use chat - you use it every day with AI, and with colleagues.
But chat alone isn鈥檛 enough for real internal workflows - we鈥檙e taking it further.
91制片厂 Chat will surface dynamic forms, tables, and charts within the chat experience, combining the flexibility of free-text with the structure of traditional user interfaces. And, of course, all securely connected to your data.
Release Schedule
| Date | Release | Features |
|---|---|---|
| Dec 2025 | Agents Alpha | Agent node (automations) |
| Feb 2026 | Agents Beta | Evals, logs, testing, stability |
| Feb 2026 | Chat Alpha | Chat interface connected to agents |
| Mar 2026 | Chat Beta | - |
| Q2 2026 | Agent Testing | Interface for validating behaviour before deployment |
| Q2 2026 | Agent Analysis | Visibility into agent actions and decisions |
| Q2 2026 | Human-in-the-loop automations | Apps/Screens inside automations |
91制片厂 has always enabled teams to build secure internal workflows. Now, we鈥檙e bringing AI into that world in a way that matches the needs of privacy-first organizations.
AI workflows, private LLMs, strict data residency, full RBAC, and multi-agent orchestration - all under your control.
If you want to help shape this future, . I鈥檇 love your feedback as we build the next chapter of 91制片厂.