CollectivIQ has introduced a broad set of upgrades to its artificial intelligence platform, adding new tools for model choice, image generation, file handling and project organisation.
The Boston-based company says the platform queries multiple large language models at the same time, including ChatGPT, Claude, Gemini and Grok, before producing a single annotated response. The approach is intended to reduce dependence on any one model and address concerns about hallucination and bias in AI outputs.
The release is CollectivIQ's most extensive product update so far. Alongside its core consensus-based response system, it adds multimodal functions and administrative features aimed at broader use across teams.
A central addition is a model selection tool that lets users choose which language models respond to a prompt. Employees can use different models for different tasks while still having the option of a consensus answer drawn from a wider group of systems.
The platform can also now generate images through the same chat interface used for text prompts. Users can request either a single-model result or what CollectivIQ describes as a combined "Best of the Best" output.
File capabilities
Another update adds payment capture, allowing individuals and organisations to manage billing and access within the platform. This is intended to simplify onboarding and expansion across a business.
File support has also expanded through retrieval-augmented generation features. Users can upload structured and unstructured files, including Word, PowerPoint, Excel, PDF, Python, JSON and plain text documents, and the system can query those materials to inform its responses.
If an unsupported image format is uploaded, the platform converts it to PNG. CollectivIQ has also added file generation inside chat threads, allowing users to ask the system to produce text outputs and downloadable files directly within a conversation.
When several models are used for the same task, the platform can merge the strongest elements of those responses into a single file, extending the same consensus logic from text answers into document creation.
Team workflow
For teams managing longer-running work, CollectivIQ has introduced a projects feature. It lets users group related chats into dedicated workspaces and attach reference documents that become shared context across those conversations.
The change is designed to make it easier for colleagues to work from the same source materials and maintain continuity across multiple exchanges. It also gives users a more structured way to manage work that extends beyond a single prompt or session.
CollectivIQ has also added what it calls a triager, a tool intended to identify user intent more accurately. The system is designed to determine whether a prompt is asking for a text response, image creation or file generation, then route or flag requests when the selected model cannot handle the task.
Positioning strategy
The upgrades come as businesses continue to test how far generative AI can be trusted in day-to-day work. Many organisations have adopted tools from more than one vendor while trying to manage accuracy, oversight, cost and staff preferences.
By placing several external models behind a single interface, CollectivIQ is positioning itself as a management and synthesis layer rather than a direct competitor to the largest model developers. Its pitch rests on the view that no single model is reliable enough for every business task.
John Davie, Chief Executive Officer, CollectivIQ and Buyers Edge Platform, framed the release around that problem.
"Organisations deploying AI are quickly realizing that fragmented, single-model tools aren't reliable for delivering accurate results. We built CollectivIQ because we lived this problem ourselves," said John Davie, Chief Executive Officer, CollectivIQ and Buyers Edge Platform.
"With CollectivIQ's new product features, we're giving teams the flexibility to choose the right model for every task, the ability to collaborate in one place, and the visibility into usage that leaders need to scale AI effectively and responsibly," added Davie.