> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kontratar.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Forecasts

> How to browse planned procurements that agencies have announced but not yet solicited, using the forecast calendar, AI Chat, Projects, and scheduled Research. 

**Forecasts** are procurements that agencies have announced but not yet released as formal solicitations. Agencies publish them to give industry advance notice of what they plan to buy.

A forecast is the earliest point at which a requirement becomes visible to you. That is both its value and its limitation. You get months of warning, but you are looking at a plan rather than a commitment.

Use the Forecasts module to:

* Find requirements before they reach the market
* Begin capability development ahead of the solicitation
* Reach contracting officers while the requirement is still being shaped
* Identify and approach teaming partners early
* Decide where to spend your business development time

<Warning>
  **Forecasts are projections, not guarantees.** Agencies change plans, delay procurements, and cancel them outright. Estimated values and expected release dates move. Check the current status with the agency before committing real capture resources to a forecasted requirement.
</Warning>

<Info>
  **Forecasts vs. Opportunities vs. Awards:** Forecasts are planned procurements not yet solicited. Opportunities are active solicitations currently posted on SAM.gov. Awards are completed procurements. The typical lifecycle is: Forecast → Opportunity → Award.
</Info>

**How to get there:** Sidebar → Click **Forecasts**

## The Forecasts workspace

Forecasts has its own rail, which replaces the main Kontratar sidebar while you are in the module.

| Rail item            | What it opens                                       |
| :------------------- | :-------------------------------------------------- |
| **New Conversation** | A new AI Chat conversation about forecast data.     |
| **Chats**            | Your previous conversations.                        |
| **Projects**         | Workspaces for organizing related searches.         |
| **Agencies**         | Forecast data filtered to one agency and subagency. |
| **Calendar**         | The forecast dashboard and calendar view.           |
| **Research**         | Questions that run automatically on a schedule.     |

<Note>
  **If you already know the Awards module, most of this will look familiar.** AI Chat, Projects, and Research work the same way in both places. The difference is what they search. In Awards they search contracts that have already been awarded. Here they search procurements that have not happened yet.
</Note>

## 1. The forecast calendar

The Calendar page is the main view of the module. It shows what is coming and when.

Three cards sit across the top:

| Card                       | What it counts                                                                              |
| :------------------------- | :------------------------------------------------------------------------------------------ |
| **Total Forecasts**        | How many forecasts are in your current results. This number changes when you apply filters. |
| **Upcoming Solicitations** | Forecasts expected to reach solicitation in the next seven days.                            |
| **Upcoming Awards**        | Forecasts expected to reach award in the next seven days.                                   |

<Note>
  **Total Forecasts reflects your filters, not the whole database.** The card is labelled Current result set. If the number does not move when you expect it to, check whether you clicked **Apply Filters**.
</Note>

### Reading the calendar

The calendar shows one month at a time, with month and year dropdowns above it. Blue dots on a date mean forecast activity is scheduled that day. Where there are more than a few, a small `+` number appears next to them showing how many more there are.

Click any date to load that day's forecasts into the panel on the right. The panel is headed with the date and a count, for example `Forecasts on July 29, 2026`, with a count of how many are scheduled.

Each forecast in that panel shows:

* **The title.** For example, `Hazen Building Renovation` or `28902443 Repair Sanitary Sewer Line at Sector Key West in Key West, FL`.
* **A description** of what the agency plans to buy. The Hazen Building entry reads: *The United States Secret Service has a requirement to redesign and renovate the security and visitor center portions of the Hazen building.*
* **Agency.** The awarding agency and the contracting office beneath it, for example Department of Homeland Security, Office of Procurement Operations.
* **Key Dates.** The milestone dates recorded for that forecast.

Click any forecast to open its full detail.

### Filtering forecasts

| Filter                  | How it works                                                                   |
| :---------------------- | :----------------------------------------------------------------------------- |
| **Filter by Agency**    | A searchable list of federal agencies.                                         |
| **Forecast Date Range** | A **From** and **To** date pair. Click either field to open a calendar picker. |
| **NAICS Code**          | A text box. Type one code, or several separated by commas.                     |

Click **Apply Filters** to run them. Click **Clear Filters** to reset the panel. The **Total Forecasts** count at the bottom of the panel updates to match your results.

<Tip>
  The NAICS box here accepts any code you type, rather than limiting you to the codes your tenant subscribes to. That means you can research a market you are not currently set up for, which is useful when you are deciding whether to move into one.
</Tip>

## 2. Agencies

The Agencies page narrows forecast data down to a single agency and, if you want, a single contracting office within it.

1. Click **Agencies** in the Forecasts rail.
2. Choose an agency from **Filter by Agencies**.
3. Choose a contracting office from **Filter by Subagencies** if you want to narrow it further.

Matching forecasts load below.

<Note>
  **`No forecast data found for the selected agency or subagency` is not an error.** It means that agency has not published forecasts matching what you selected. Try a broader subagency, or check a different agency.
</Note>

## 3. AI Chat

AI Chat lets you ask questions about forecast data in plain English instead of working through filters.

It works the same way as AI Chat in Awards, with one extra control: you can scope your question to a single agency before you ask it.

To ask a question:

1. Click **New Conversation** in the Forecasts rail.
2. Optionally, click **Select agency** beside the prompt box and choose one. The list has a search box, so you can type instead of scrolling.
3. Type your question.
4. To include a supporting document, click **Attach** and choose the file.
5. Click the arrow button to send.

Questions that work well:

* Which forecasts involve software license renewals in the next 12 months?
* Show me DoD forecasts related to cloud migration
* What companies currently hold contracts related to this forecasted requirement?
* Tell me all I need to know about this contract
* Which upcoming procurements match NAICS 541511?

<Tip>
  Using the agency picker is more reliable than naming the agency inside your question. The picker limits the search directly. A name typed into your question is something the AI has to interpret first, and agency names are long and easy to get slightly wrong.
</Tip>

### Continuing a previous conversation

Click **Chats** in the Forecasts rail to see your earlier conversations. Recent ones also appear under **Recents** at the bottom of the rail. Click one to reopen it with its full history and carry on where you left off.

## 4. Projects

Projects help you organize related searches. A project holds instructions and supporting files, and every chat you create inside it uses them as context.

1. Click **Projects** in the Forecasts rail.
2. Click **New Project**.
3. Type a **Project Name**, and a **Description** if you want one.
4. Click **Create Project**.
5. Open the project and add your instructions and any supporting files.

For example, a project with the instruction `Only focus on Department of Homeland Security forecasts for building renovation work` will apply that filter to every question you ask inside it. You would not need to repeat it.

<Note>
  **Projects do not cross between modules.** A project you create here searches forecast data and appears only in the Forecasts rail. A project created in Awards searches award data and stays there. One project cannot cover both.
</Note>

## 5. Research

Research automates recurring questions so you do not have to ask the same thing every week.

1. Click **Research** in the Forecasts rail.
2. Click **Create Research**.
3. Enter your question, a name you will recognize later, and how often it should run: Daily, Weekly, Monthly, or Yearly.
4. Save it.

Kontratar runs the question on your schedule and files each result under the date it ran. Because runs are stored separately rather than overwriting each other, you can open one research item and read several dates in sequence to see what changed.

The table lists each scheduled question with its name, the prompt, the interval, and the date. The search box above filters the list.

<Tip>
  Forecast data moves more slowly than award data. Agencies update their forecasts periodically, not daily. A monthly interval usually gives you a meaningful comparison. A daily one mostly gives you the same answer thirty times.
</Tip>

## 6. Using forecasts strategically

The window between a forecast being published and the solicitation being released is the only period in which a vendor can realistically influence a requirement. Once the solicitation is out, the requirement is fixed.

| When                                 | What to do                                                                                                                                |
| :----------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------- |
| **As soon as you find the forecast** | Check it against your capabilities. Add it to [Pipeline](/Version-2/Pipeline-Management) at Stage 1 or 2.                                 |
| **6 to 12 months out**               | Start capability development. Work out which gaps you would need a teaming partner to fill.                                               |
| **3 to 6 months out**                | Approach the contracting office. Use [Awards](/Version-2/Awards) to find who holds the related work now and how strong their position is. |
| **When the solicitation appears**    | The requirement moves to Opportunities. Advance your pipeline record to Stage 3 and make the bid or no-bid decision.                      |

For example, if you find a forecast in March for a geospatial support services contract expected to be solicited in October, you have seven months. That is enough time to build a teaming relationship, talk to the contracting office about what they actually need, and look up the incumbent's award history in the Awards module. A competitor who first sees the requirement in October has none of that.

## Related topics

* [Awards](/Version-2/Awards). Completed procurements. Use these to identify who currently holds work related to a forecast.
* [Pipeline Management](/Version-2/Pipeline-Management). How to track a forecasted requirement through the 8-stage capture lifecycle.
* [Campaigns](/Version-2/Campaigns). How to reach contracting officers while a requirement is still in the planning phase.

**Parent topic:** [Kontratar v2.0 Documentation](/Version-2/whats-new-v2)