Skip to main content

How to run a sourcing case with Speya Auto

Speya is your AI assistant for sourcing. You describe what you need and it calls the right tool for the job: Discovery to find suppliers, Enrich to research them, Engage to email them, and Monitor to track them over time.

This guide covers Speya Auto, the hands-off way to run a sourcing case. In the normal step-by-step flow, a search builds one discovery plan and runs it, then stops. Speya Auto removes that back-and-forth: from a single request, Speya runs several searches in parallel, keeps going on its own, and sorts what it finds into tiers. You give Speya your case and step away, and it comes back with a ranked shortlist. Use it when your case is well briefed and you want a finished, tiered list without steering every step.

In this article

  • Run a sourcing case with Auto

  • Understanding your results

  • When is your Auto case done?

  • Best practice tips

  • Prompt library

  • FAQ


Run a case with Auto

1. Switch the search mode to Speya Auto. In the search screen, select Auto in the bottom-left.

2. Describe your case with as much context as you can. The more context you give, the better Speya scopes the run, so paste the background and goal, not just the product. A good starting prompt is:

“I’m sourcing [product] for [application]. Find [manufacturers / distributors] in [region], not [what you don’t want]. Today we buy this from [current supplier], but [they’re too expensive / we’re single-sourced / the part is discontinued].”

If your product is technical and a document describes it better than a sentence can, attach the spec and say “Start discovery.”

3. Answer a few follow-up questions. Speya runs an initial research pass, then asks a few questions to set the boundaries of the case, such as which product variant you mean or which region to source from. Pick one of the suggested answers or type your own.

4. Let Speya Auto run the case from start to finish. Once the case is scoped, Speya runs the whole case on its own as background tasks, so you can move to other work and come back to a finished shortlist. On its own, it:

  • Runs several supplier searches in parallel. It builds one or several discovery plans that each describe your product or service a different way and runs them at the same time, so it reaches suppliers that a single wording would miss.

  • Keeps looking until the list settles. It reviews the suppliers it finds and decides whether to adjust a plan or add another search to find more that fit your case, and it stops when new searches stop adding suppliers.

  • Builds the enrichment columns. It has already researched your case, so it knows what to look at, and it creates the enrichment columns that matter without you having to specify them.

  • Sorts suppliers into tiers. It reviews the full table of suppliers and their research, sorts them into three tiers, and summarizes the top-tier suppliers in the chat so you get a quick read on the most interesting ones.

  • Removes duplicates. It clears out any supplier that appears more than once.

  • Offers next steps. When the list is ready, it gives you options for what to do next: email your top-tier suppliers and find their contact details, deep-research a supplier, or create a spreadsheet.

5. Add your own enrichments while Speya works. You don’t have to wait for the case to finish. While Speya is searching, enriching, and tiering, you can chat with it to add enrichment columns or ask for extra research, for example “Also research [enrichment] across the whole list.”, and it runs them in parallel with its own work.

6. Read the tiered suppliers. Speya sorts suppliers into Tier 1 to Tier 3 and summarizes each supplier’s position based on what your case needs. In the Tier column, click the info icon on a supplier to read its reasoning, and hover over the Tier column header to see the instructions Speya used to tier. Speya builds the tiers from what it has learned about your case so far and the data it has access to, so they are not fixed. To change how the tiers are decided, ask Speya to re-tier on your own criteria. The tiers stay live after the run, so they update as you add enrichments, run more discoveries, or ask Speya to re-tier.


Understanding your results

Your results are a table: each row is a supplier Speya found, each column is an enrichment it researched, and the Tier column ranks them. If you only ever use Auto, this table is where your sourcing work happens, so it is worth knowing how to read it.

  • Suppliers (rows). Each row is one supplier. Open it to see everything Speya has gathered about it in one place, which fills out as the case runs.

  • Enrichment columns. Each enrichment is one thing Speya researched about every supplier, such as certifications, annual turnover, or whether they serve your industry. Speya chooses the enrichments that fit your case, and you can add more at any time. For the full detail on enrichments, see How to research suppliers with Enrich.

  • Cell states. A cell marked Match means Speya found evidence for that enrichment, with a source link to where on the web it found it. A cell marked Run means it is still searching. A cell marked not run means the column belongs to a different discovery plan and has not been applied to that supplier yet; use Research missing to fill every cell.

  • Sources. Everything on the list is source-backed and traceable, and Speya does not guess. When it cannot confirm something, it says so rather than filling the gap.

A Match confirms that a public trace of something exists, not the live detail behind it. Things like current price, real stock, and an exact dimension or finish change too often to read reliably from public sources, so treat those as questions for the supplier rather than facts on the list. That is what the RFI step is for.


When is your Auto case done?

Speya works toward a finished list on its own, so “done” is less a moment you trigger than a state the list reaches. It keeps checking whether the pool is stable, meaning no new suppliers turned up in the last round of searching, adjusts the tiers, and tells you when more searching is unlikely to add anything. At that point your tiered list is ready to act on.

It is normal for a few rows to stay thin where a supplier simply does not publish a detail; that is a gap in public information, not a failure of the list, and it is often a good question to put to the supplier directly. For more on how enrichments work, see How to research suppliers with Enrich.

When the tiers look right, ask “Create an executive summary of the top suppliers for this project.” Speya builds an Excel spreadsheet of the top suppliers that you can share with colleagues who haven’t seen the list.

When the list is ready, Speya offers you the natural next steps. The most common is contacting your top-tier suppliers to confirm the specifics public research can’t. See How to send RFIs with Engage. To keep your chosen suppliers tracked for changes over time, see How to import and track suppliers with Monitor.


Best practice tips

  • Give Speya a real brief, not a one-liner. Speya scopes the whole run from the context you give it, so paste the email, the spec, or the background rather than a single line. Naming your current supplier and why you are looking gives it a benchmark for what “good” means in your case.

  • Answer the scoping questions with care. The questions Speya asks up front, such as variant and region, shape every discovery plan and column that follows. Precise answers here are the cheapest way to get a precise list.

  • You don’t need the perfect prompt. Supplier information is spread across the web, trade data, and registries, and it is described inconsistently, so describing your product one way reaches a different set of suppliers than describing it another way. Speya searches several framings of the same product in parallel, one broad, one with your must-haves, and one simplified, so your coverage comes from the range of searches rather than from landing on a single right phrase.

  • Expect Auto to take longer to load than step-by-step. Speya runs several plans and evaluates the whole table, so it takes longer to reach a list. The trade is that it runs hands-off in the background while you do something else.

  • Keep working on the same list. Lists accumulate and columns persist, so add enrichments and refine in place rather than starting a fresh case each time you have a new question.


Prompt library

Every prompt here is something you type to Speya. They’re grouped by where you are in an Auto case, with a description of what each does.

Start an Auto case

  • “I’m sourcing [product] for [application]. Find [manufacturers / distributors] in [region], not [what you don’t want]. Today we buy this from [current supplier], but [they’re too expensive / we’re single-sourced / the part is discontinued].”
    Starts an Auto case from a full briefing. Naming your current supplier and the reason you’re looking gives Speya a benchmark for what “good” looks like.

  • “Start discovery.” [attach: technical specification PDF]
    Starts the case directly from a spec. Best when the product is technical and the document describes it better than a sentence can.

Steer the research while it runs

  • “Also research [enrichment] across the whole list.”
    Adds an enrichment to every supplier, while Speya works or after it finishes.

After the run

  • “Create an executive summary of the top suppliers for this project.”
    Builds an Excel spreadsheet of the top suppliers, ready to share with colleagues who haven’t seen the list.


FAQ

How is Auto different from a normal search?
A normal search builds one discovery plan and runs it, then you add enrichments, evaluate, and shortlist manually. In Auto, Speya runs several plans in parallel from one case prompt and a few follow-up questions, keeps iterating until it stops finding new suppliers, then evaluates and tiers the list for you. That gives you a sharp shortlist you can act on directly: send an RFI, or create an executive summary to share with colleagues.

Do I have to stay at my computer while Speya runs the case?
No. Speya runs the case as background tasks, so you can step away and come back to a finished shortlist.

Why does Auto take longer than doing it manually?
Because Speya does more without you: it runs several supplier searches in parallel, researches the enrichments, evaluates the full table, removes duplicates, and tiers the list. None of it needs your attention while it runs, and when it is done you get the most relevant suppliers presented for you.

Can I still control what Speya researches?
Yes. Add prepared enrichments at any time, or prompt your own enrichment columns alongside Speya’s while it works.

Can I change the tiers?
Yes. Change a supplier’s tier manually by clicking on it, or ask Speya to update one or several suppliers at once.

What happens if a search errors or finds nothing?
Speya retunes and tries again. It closes a plan that stops producing, evaluates what it has, and comes up with a new discovery plan, looping until the pool stops growing.

When should I use Speya Auto instead of step-by-step?
Use Auto when your case is well briefed and you want a finished, tiered shortlist hands-off. Use step-by-step when you want to control each step, or when you are still working out how to frame the case and want to see each plan before it runs.

Did this answer your question?