Public Roadmap

Milestone: Users are recommended AI-generated models

Ability to add labels (or tags) to forecasts and collections

We need a better way to organize collections. After adding many collections it becomes difficult to find what you’re looking for from the main collections page.

Search matches w/ exact model 1st

If you search an exact model name, for example “Engineered Enzymes For Biomanufacturing” it would be great if that appeared first. Currently it is appearing 4th. Another consideration is highlighting the result that is an exact match to the search term.

Milestone: Koi facilitates impact forecast task management

Tutorials for custom modeling

A step-by-step tutorial showing the in-app process and scientific logic for creating and interpreting an entirely custom model.

Add Lifecycle Sub-phases

Expose the lifecycle subphase data through the UI for end users.

Show additional impact and risk factors

Include a page with solution details that outlines other impacts of the solution (e.g. waste, water, plastic) and potential risks (e.g. impact to workforce)

Title of market capture in collection summary

Once the refactor is complete and we have titles for each market capture ID, it would be suppppper helpful to see in the exec summary of a collection.

Adapt search bar to accommodate multiple models with same name

This is not a bug but more of a personal observation/suggestion. When searching using the “rooftop” keyword, multiple models appear with the same name. After selecting “preview,” I recognize they are slightly different (in terms of market capture, description, etc.), and I get that they’re probably generated based on different company IPs. From a UX perspective, it would save time and effort for a user to be able to distinguish between these models more easily. For instance, if there was an extra column listing those key differences, or if models would somehow be pooled into one master “Rooftop PV Systems” category (this one is especially helpful when a user is searching for something not solar PV related, yet getting so many irrelevant results).

Add new member to group who does not have account

When adding users to a group, if the user does not yet have a Koi account we need to either 1) let them be added anyway or 2) provide a useful error message. Right now it just doesn’t let you add (which is fine) but no useful message to user.

Milestone: Public API to Koi Data Lake

Expose Koi Data Lake over public API. Get in touch if you’d like to be a beta tester, support@koi.eco

Koi engine 2.0

Koi engine 2.0 will revise the approach to calculating solution impact. Rather than assuming an intervention in the baseline value chain, by construct default in 1.0, it will attempt to model the solution’s distinct value chain based on credible research.

In-app QA process

Limitation addressed: It is unclear to Koi users what data are verified in a model and what data are autogenerated. If they modify data it is up to them to remember what they changed and communicate it. If they are more comfortable working in a spreadsheet, they have no way of doing so. Develop tooling for Koi studio users to understand and manage the QA status of their model. Data should be appropriately tagged at the parameter level for communications of model quality and verified/unverified information. QA tooling to include streamlined “high-level” model refinement options as well as deep refinement using an exported Review.xlsx file that can then be reingested from an updated Google Sheet.