Import
Upload a CSV or an Apollo.io export. The wizard previews headers, fuzzy-maps them to prospect fields, validates, and deduplicates on email. Large files run as a background job.
Forge module · Valkyrie Assist
AI-enriched prospecting that graduates into your CRM.
Import a list, enrich every row with AI columns you define, qualify with the reasoning attached, and promote the winners into your The Viking Stack CRM with mapped fields and tags. Prospects are pre-CRM leads: they stay out of your contact database until they earn a place in it.
Built for the same teams that use Forge AEO: sellers whose buyers research by spec, standard, and quote terms, and whose outreach needs to sound like it knows that.
Import, enrich, qualify, promote.
Four steps, each with a review point. Nothing reaches the CRM until a person, or a rule you wrote, says it should.
Enrich
Define AI columns with your own prompts, context fields, and output format. Run one prospect instantly or queue the whole list with the cost estimated first.
Qualify
Sort, filter, and segment by status, source, tags, and AI results. Every value carries the model's reasoning so a rep can check the call in seconds.
Promote
Push single or bulk selections into the CRM with mapped fields, additive tags for workflow triggers, and an optional note with the reasoning. Failures surface on the row.
What Valkyrie Assist gives your team.
CSV import wizard
Upload, parse, auto-map, validate, import. Existing prospects merge by email without wiping earlier AI results; rows with no identifier are rejected with a per-row report.
AI enrichment columns
Prompt templates with prospect fields, your domain, dates, and conditionals. Outputs as text, boolean, number, or JSON, each stored with its reasoning and execution history.
Provider choice
Claude, OpenAI, Gemini, Perplexity, or xAI per column. Run on the platform allowance or bring your own keys; the UI shows which lane applies before you spend.
Knowledge-base profiles
Named context documents about your products, ICP, and positioning that any column can inject, so enrichment sounds like your company, not a generic model.
Segments and review
Saved segments over status, source, tags, and AI results, plus a review queue for the prospects a human should look at before promotion.
Field mapping to your CRM
Map prospect fields and AI columns to CRM custom fields, list the fields that exist, and create the missing ones from the same screen. Sensible defaults ship enabled.
Cost control
Every bulk run shows its price before it starts.
AI enrichment is only useful if the bill is predictable. Bulk runs are queued as jobs with an estimated cost, platform-lane runs reserve spend against your account allowance, kill switches stop a runaway job, and every execution records tokens used and cost so finance can reconcile it.
The task monitor shows progress for imports and bulk enrichment alongside your AEO audits and publishes, so one screen answers "what is Forge doing right now."
Estimated cost on every bulk enrichment job before it runs.
Spend reservations and kill switches on the platform lane.
Per-execution tokens and cost recorded on the prospect.
Bring-your-own keys per provider, stored as secret references, never displayed.
No fixed row cap; large lists run as background jobs.
Connected to Journeys & Pulse and Apollo.io.
Inbound form submissions can create or update prospects automatically, and a linked prospect's detail view shows the web journey behind it. On the outbound side, Apollo.io exports import directly, so research lists and inbound requests meet in one queue before either reaches the CRM.
Form submissions feed prospects with source and attribution attached.
Journey activity panel on the prospect detail when a visitor is linked.
Apollo.io lists imported as a first-class source.
Promotion links the prospect to the CRM contact and its journey identity.
Built on the same platform rules.
Account-scoped everywhere, row-level security on every table, Forge design system in light and dark, and the same connections and task monitor the rest of Forge uses. Valkyrie Assist is a module, not a bolt-on.
Valkyrie Assist questions.
Why keep prospects out of the CRM?
Research lists are noisy. Valkyrie Assist holds prospects as pre-CRM leads so enrichment, qualification, and cleanup happen before anything touches the contact database. Promotion is graduation: only promoted prospects become CRM contacts, and re-promoting updates the same contact rather than creating a duplicate.
Which AI providers can run enrichment columns?
Claude, OpenAI, Gemini, Perplexity, and xAI. Each column names its provider. Accounts can run on the platform allowance or bring their own per-provider keys, and the builder shows which lane a column will use before you run it.
How do we control AI spend?
Single-prospect runs execute immediately. Bulk runs are queued with an estimated cost shown up front, platform-lane runs reserve spend against your allowance, and kill switches stop a runaway job. Every execution records tokens used and cost.
What happens on promote?
The contact is upserted with your mapped standard and custom fields, mapped AI-column values land in CRM custom fields, tags are added (never overwritten), and an optional note carries the AI reasoning. Delivery goes through the same idempotent, retrying queue Journeys & Pulse uses.
Does it pull from or sync with the CRM?
No. Valkyrie Assist is one-directional by design. The Viking Stack remains the sole system of record for contacts, and nothing reaches it except through the promotion flow.
Is Valkyrie Assist a separate app?
No. It is a module inside Forge alongside AEO & Agentic Storefronts and Journeys & Pulse, sharing the same account, connections, design system, and task monitor.
Turn research lists into a CRM your reps trust.
Book a free AI Visibility Audit. We will map how your team prospects today and where AI enrichment, review, and promotion fit before anything touches your contact database.