How to Create a Project in Seedli: A Step-by-Step Guide

From any screen in Seedli, open the Projects dropdown in the top-right header. The button labelled Add project sits at the bottom of the dropdown, below your existing projects. Click it to start the setup flow.

On the Starter plan you can run up to three projects in parallel. On Scale and Enterprise the project caps go higher, and you can add more individual projects as add-ons. If your plan is at its cap, Seedli prompts you to upgrade or add a project before the setup flow opens.

The first screen tells you what you’re about to build and how long it takes. You describe your brand and audience in 5 to 10 minutes. Seedli runs the analysis across AI models. Results land in about 30 minutes.

The three cards below the headline preview the output you’ll get back: who AI recommends in your category, what eliminates brands before buyers reach out, and where your brand goes dark. These are the questions the analysis is built to answer. Read them. They’re the lens for every choice you make in the next three steps.

Click Start setup when you’re ready. Nothing is locked in yet; you can go back at any point in the flow.

Three fields, only one required. Company name is the brand exactly as buyers and AI models would refer to you. Use the version that appears in the market, not a legal entity. If you’re known as Acme but registered as Acme Industries Holding AB, type Acme.

Company website is optional but worth filling in. Seedli uses it to verify your brand and domain in AI-generated citations, which improves how the analysis attributes mentions back to you when AI cites your domain rather than your brand name.

Project name is also optional. If you plan to run one project ever, skip it. If you plan to run multiple (different markets, different audiences, different verticals), pick a name now that scales. A name like UK Mid-Market 2026 tells you and your team what the project covers the moment you see it in the dropdown.

This is the screen that decides how sharp the analysis comes back. Seedli feeds your description to AI models to extract how they categorise your space. A generic description produces a generic map. A specific description produces a sharper competitive picture.

The five inputs:

Click Draw the map when you’ve filled in all five required fields. The button turns from muted teal to solid teal once enough is filled in.

Seedli runs the first-pass analysis twice and compares the two reads. The comparison is the bias check: two AI runs against the same description should produce a similar market frame. When they diverge, the analysis uses the overlap rather than either single read.

The screen shows three states: First read, Second read, Comparing. The whole step takes 20 to 40 seconds.

Step 3 shows you the market frame the AI derived from your Step 2 inputs. Everything on this screen is editable. Small fixes here make a big difference downstream, because the full 30-minute analysis runs against the values you confirm now.

The Clarity Score in the top right is the AI’s confidence in how unambiguously your market is framed. 80% and above is ready to ship. Below that, the AI is signalling that something in the frame is ambiguous enough to dilute the analysis.

The fields you review, in order:

Near the bottom of Step 3, you’ll see a panel titled One detail can sharpen your analysis. The AI reads your Step 2 description, identifies the ambiguity that’s holding the Clarity Score below 100%, and proposes a rewrite. Two buttons: Apply and Skip.

Apply when the proposed wording is sharper than yours, or when the AI identifies a real positioning ambiguity you hadn’t spotted. Skip when the AI’s rewrite is more generic than your original, or when the ambiguity is actually structural to your product (you span two categories on purpose) rather than a wording problem.

The Clarity Score is diagnostic, not a target. A score of 88% can be a ceiling rather than a defect. If the AI’s reasoning describes your product accurately as “straddling two buying categories,” and that’s a deliberate positioning choice, the score won’t move regardless of how you rephrase. In that case, skip.

The bottom-left of Step 3 has an ← Edit input link. It takes you back to Step 2 with all your values still filled in. Change anything you want, then submit again. Seedli re-runs the read-twice-and-compare analysis against your new inputs and shows you a fresh Step 3.

Use this when the market frame on Step 3 doesn’t match how your buyers actually describe your category. The most common reasons to go back:

• The anchor label is too broad (the AI grouped you with an adjacent category that buyers don’t actually conflate with yours).
• The competitive roles include a category you don’t belong in, or are missing one you do.
• The market topic uses a phrasing your buyers wouldn’t search for.
• The Clarity Score is below 80% and the rewrite suggestion doesn’t explain why.

You can re-run any number of times before launching the full analysis. Once you click Yes, Analyse This Market, the 30-minute analysis runs against whatever the market frame looked like at that moment. The frame is locked for that run.

If the analysis comes back and the frame turns out to have been wrong, the fix is to set up a second project with the corrected frame rather than editing the existing one mid-run. Each analysis is a self-contained measurement of your market under one frame; mixing frames mid-run produces data you can’t compare.

Clicking Yes, Analyse This Market launches the full analysis pipeline. The screen shows you which stage you’re on and which stages are queued behind it. The top block runs first, then the queue runs in sequence.

The full pipeline runs roughly thirty minutes end to end. You don’t need to keep the tab open. Close it, work on something else, and check back when you have a moment. Seedli sends an email when the analysis is complete, and the dashboard updates the moment each stage finishes, so partial results are visible as they land.

The first stages produce the competitive map and the elimination triggers. The later stages produce the buyer-signal data that powers the Content Plan. Once the pipeline finishes, the Content Plan section of your project becomes the most actionable view: it ranks the specific gaps between how AI currently talks about your brand and how it should, and turns each gap into a content brief you can hand to a writer.

If you want to set up a second project for a different market or audience while the first runs, you can. The two analyses are independent and run in parallel.

Once the analysis pipeline completes, the first time you open the project’s Consideration view (the page where the Content Plan lives) Seedli prompts you to pick a provider type. This is a one-time gate: you can’t see consideration data or the Content Plan until you choose.

The list isn’t generic. It’s extracted from your own market analysis: the categories the AI actually identified in your space, each with a short positive description of what it means and an explicit exclusion clause for what it is not. The top options carry a Primary tag (most likely fit). The rest are tagged Secondary (less likely but possible).

The choice shapes how the rest of the analysis personalises its recommendations. Pick the wrong category and the comparison set, opportunity scoring, and positioning insights all skew toward the wrong peers.

How to choose when you’re not exactly sure

1. Read the positive description, not the label. Labels like Platform or Suite can mislead on their own. The sentence under each name is the disambiguator. Pick the one that describes what you’d actually say about your product at dinner.
2. Use the exclusion clause as a check. Each option has a “Not a…” line. If the exclusion rules out something you genuinely are, that’s the wrong category. If it rules out things you clearly aren’t, that’s a confirming signal.
3. When two options feel close, pick the one your competitors would pick. Provider type decides who you get benchmarked against. You want to be measured against your real comparison set. Look at the competitor list you locked in Step 3, and ask which option those brands sit under.
4. You can change it later. Provider type lives in project settings. If your first choice turns out wrong, switch it, and the next refresh of the Content Plan reflects the new framing.