Glossary
Plain-language definitions of every term you will encounter in Ken AI and cold-email outreach - for users learning the platform and for support agents diagnosing issues.
Overview
This glossary covers campaign concepts, deliverability terms, AI pipeline stages, and platform-specific vocabulary used throughout the Ken AI dashboard. Refer to it whenever a label, metric, or status is unclear. Support agents should treat these definitions as the authoritative baseline when investigating tickets.
There is no single page in the app for this content - terms appear across Campaigns, Email Delivery, Replies, and Settings. Use this as a companion reference while navigating those sections.
How to use this glossary
Search or scroll to the term you need. Each entry gives a plain-language definition followed by notes on where the concept appears in the app, if it has a dedicated UI location. Terms that are aliases for other terms (such as Signal and Trigger) point to the canonical entry.
If you are an AI support agent, use the Troubleshooting & debugging section first - it maps symptoms to root causes directly. The term definitions give the underlying concepts needed to reason about those causes.
How it works
Ken AI's vocabulary is organized in layers that follow the campaign lifecycle:
- Targeting layer - Audience, ICP, Segment, KenSearch, and Contact define who receives outreach.
- Enrichment and verification layer - Enrichment, Verification, Bounce, and Catch-all domain determine whether a contact is reachable.
- AI pipeline layer - AI flow (Qualify, Segment, Personalize, Rewrite), Prompt, Sequence variables, and Personalization describe how Ken writes and scores the emails.
- Sending layer - Sender, Sender config, Campaign mode (Batch/Trigger), Email sequence, and Step control when and how emails go out.
- Deliverability layer - DKIM, SPF, DMARC, DNSBL, Warmup, Deliverability score, Placement, and Reconciliation describe the health of the sending infrastructure.
- Feedback layer - Intent signal, Approval, and Workspace-level settings govern review and re-engagement.
Most support issues trace back to a break between two adjacent layers - for example, a contact that cleared enrichment but failed verification, or a campaign with a valid sequence but no sender config attached.
Before you start
No special role or plan is required to read this article. Some features described here (such as Trigger campaigns or AI flow overrides) are available only on certain plans. Where that matters for diagnosis, it is noted under the relevant term.
Term reference
A/B variant
A named alternative version of a step in an email sequence. You can create multiple variants on the same step and assign a percentage split across them. Ken runs the AI personalization pipeline on each variant independently before launch. Analytics break down opens, replies, and clicks per variant so you can measure which copy performs better.
In the app: sequence editor under Campaigns - Actions - Sequence. Each step shows a "+" button to add a variant.
AI flow
One of the four stages in Ken's AI pipeline: Qualify, Segment, Personalize, and Rewrite. Each flow runs independently, has its own prompt, its own model setting, and its own cost meter. When you test a prompt in the app, you are testing one flow at a time.
Qualify - decides whether a contact fits your ICP. Defaults to "qualify when uncertain" to avoid dropping real prospects.
Segment - assigns each qualified contact to exactly one segment using chain-of-thought reasoning.
Personalize - writes the personalized variable content (first lines, PS lines, subject lines) for each contact.
Rewrite - applies style rules and runs the adversarial evaluation loop. A line must score 4.5 out of 5 and pass all hard rules before it sends.
Approval (client approval)
A magic-link workflow that lets a client sign off on a campaign before launch. Ken builds a 20-contact sample through the full AI pipeline and generates a shareable link. The client sees the real audience, sample leads, and the actual AI-written emails each lead will receive. The CSM can watch edits arrive in real time before the client clicks Submit.
In the app: Campaigns - Actions - Ask Approval.
Audience
The targeting criteria that define who goes into a campaign. An audience is built from filters (job title, industry, company size, location, and more), KenSearch imports, or CSV uploads. Multiple audiences can feed one campaign.
Blocklist (do-not-contact)
A per-client suppression list of email addresses and domains. Any address on the blocklist - or on any subdomain of a blocked domain - is rejected before enrichment and is never sent an email. The blocklist is kept in sync with the sending platform automatically.
Bounce
An email that was rejected by the recipient's mail server and never delivered. A hard bounce means the address is permanently invalid. A soft bounce is a temporary rejection (full mailbox, server unavailable). Ken's triple-verification step is designed to eliminate hard bounces before launch by verifying every address through multiple providers.
Campaign
The primary unit of outreach in Ken AI. A campaign ties together an audience, one or more segments, an email sequence, AI personalization settings, sender inboxes, and lifecycle state. A campaign moves through statuses: Draft, Building, Ready, Active, Paused, and Completed.
Campaign mode
Controls how a campaign fires sends. The two modes are:
Batch - all contacts receive emails on a schedule. Sends go out over time to the full list.
Trigger - sends fire one at a time, one per contact, the moment an intent signal fires (a job change, a funding round, a reply classified as Interested, and others). Trigger mode requires a signal rule to be configured.
Catch-all domain
An email domain configured to accept mail addressed to any username, whether or not a real mailbox exists. Because the server says "yes" to everything, standard verification cannot confirm an address is real. Ken runs a dedicated four-provider verification race on catch-all addresses to determine whether each specific address is genuinely deliverable.
Config shard / sender config
The assignment of a pool of sender inboxes to a campaign or segment. Ken uses Hamilton apportionment (the same math used for US congressional seats) to allocate leads across inbox pools, ensuring the total adds up exactly. A campaign will not launch if no sender config is attached.
Contact
A single person in Ken's system, identified by a normalized identity (email address, name, company). The same human appearing under different spellings is deduplicated to one contact record before entering any campaign.
CSM (Customer Success Manager)
A Ken team member who manages client campaigns. CSMs use the platform to build, launch, and monitor outreach on behalf of clients. In the context of the agent MCP layer, "CSM" refers to the operator role with full platform access.
Deliverability score
A 0-100 daily health score computed for each sending inbox from five weighted signals: placement test results, bounce rate, spam complaint rate, DNSBL blacklist standing, and DNS health (SPF/DKIM/DMARC). Missing signals are dropped and the remaining signals are renormalized rather than counted as zero.
In the app: Email Delivery - Inboxes. Each inbox row shows its current score.
DKIM (DomainKeys Identified Mail)
A cryptographic signature added to every outbound email. The receiving mail server checks the signature against a public key published in the sending domain's DNS. A missing or broken DKIM record causes deliverability failures. Ken generates per-domain 2048-bit DKIM keys and syncs them live into each sending node.
DMARC
A DNS policy record that tells receiving mail servers what to do with messages that fail SPF or DKIM alignment. Ken sets new sending domains to DMARC quarantine from day one (not the common but riskier p=none) because all sending is controlled, and any unaligned mail is treated as a bug.
DNSBL (DNS-based blocklist)
A publicly maintained list of IP addresses or domains known to send spam. Mail servers check incoming mail against these lists and may reject or deprioritize mail from listed senders. Ken monitors every sending IP and domain against 60+ blocklists continuously.
Email sequence
An ordered series of email steps with configurable wait times between them. Each step can have one or more A/B variants. Ken translates sequences into the sending platform's exact wire format before launch, including wait days, variable tokens, and spacing anchors.
Enrichment
The process of finding and verifying a contact's email address and filling in missing profile data (company, title, location, technology stack, LinkedIn URL). Ken runs enrichment in a parallel pipeline that averages 2-4 minutes per contact, using multiple providers in a priority order.
ICP (Ideal Customer Profile)
The definition of your target buyer: job titles, industries, company sizes, locations, and other criteria that describe who is most likely to buy. Ken's AI qualification step checks each contact against the ICP and defaults to "in" when the evidence is ambiguous.
Inbox placement test
A test that sends real emails from client inboxes to seed mailboxes at major providers (Gmail, Outlook, Yahoo, and others) and reports whether those emails landed in the inbox, promotions tab, or spam folder. Ken staggers placement tests to look organic and runs weekly tests automatically. On-demand tests are available per inbox.
In the app: Email Delivery - Inboxes - select an inbox - Placement Tests.
Intent signal (trigger signal)
A behavioral or data event indicating that a contact may be ready to buy. Examples: a job change, a promotion, a company raising funding, a hiring surge in a specific function, or a reply classified as Interested. Signals are computed from the KenSearch index and re-evaluated monthly. Trigger campaign mode fires one personalized send per signal event.
In the app: Campaigns - Signals tab.
KenSearch
Ken's proprietary B2B contact database, built from a 280M+ person index refreshed monthly. You can search and filter by title, industry, company size, location, seniority, signals, and dozens of other fields. The index re-sweeps for intent signals on the first of each month and auto-sources matching contacts into live campaigns.
In the app: Search in the main sidebar.
Lead
A contact that has been imported into a campaign and is moving through the enrichment, qualification, and outreach pipeline. "Lead" and "contact" are used interchangeably in most parts of the app, but "lead" typically refers to a contact in the context of a specific campaign.
Personalization
The AI-written variable content added to each email for a specific contact - opening lines that reference the prospect's company, role, or recent activity. Ken scrapes each prospect's company website before the AI writes a word. Every personalized line is scored by an adversarial evaluator and rewritten until it reads like a human wrote it.
Placement (inbox placement)
Where an email lands in a recipient's mailbox: primary inbox, promotions/updates tab, or spam/junk. Placement is distinct from delivery. An email can be delivered (accepted by the server) but land in spam. Placement tests measure this before real contacts receive mail.
Prompt
The instruction text sent to an AI model to produce output for a specific flow (qualify, segment, personalize, rewrite). Prompts are stored and versioned in a prompt registry. You can edit prompts and test them live against real sampled contacts before saving.
In the app: Campaigns - AI tab - select a flow.
Qualification
The first AI flow. It reads each contact's profile, company website content, and enrichment data and decides whether the contact fits the ICP. A contact that fails qualification is excluded from the campaign and does not receive an email. The default behavior is to qualify when in doubt - dropping a real prospect costs more than one extra email.
Reconciliation
The process of comparing Ken's internal state against the sending platform's state and resolving any differences. Reconciliation runs automatically after any edit to a live campaign (sequence changes, prompt updates, new contacts) and can be triggered on demand.
Segment
A named sub-group within a campaign. Each segment has its own targeting filters, its own personalization prompts, and its own email sequence. Multiple segments in one campaign allow different messaging for different ICPs while keeping the campaign in one place. One segment is always the default.
Sender (inbox)
A sending email address, for example [email protected]. Senders are attached to campaigns via sender configs. Ken manages the warmup, DNS health, and deliverability monitoring for every sender. Inboxes are never included in a campaign launch if their warmup age is below the configured minimum.
Sequence variables
Placeholder tokens in email copy that are replaced with contact-specific values at send time. Single-brace tokens like {firstName} resolve instantly from the database. Double-brace tokens like {{openingLine}} are written by the AI personalization flow and cost tokens. The distinction matters for cost estimation.
Signal
See Intent signal.
SPF (Sender Policy Framework)
A DNS record that lists the IP addresses authorized to send email on behalf of a domain. Receiving servers check SPF to verify the sender is legitimate. A missing or misconfigured SPF record can cause mail to be rejected or marked as spam. Ken automates SPF record creation and monitors it continuously.
Step (sequence step)
One email in a multi-step sequence. Each step has a subject line, body copy, a wait time (in days) before it sends relative to the previous step, and optional A/B variants. Steps are numbered in the sequence editor and appear in analytics as Email 1, Email 2, and so on.
Trigger
See Campaign mode (Trigger) and Intent signal.
Verification (email verification)
The process of confirming that an email address is deliverable before sending. Ken triple-verifies every address: the enrichment provider's verdict is treated as provisional (finders are unreliable), and every address is re-checked independently. Once any provider marks an address Invalid, it stays Invalid permanently - even across future campaigns.
Warmup
The process of gradually increasing sending volume from a new email address to build a positive reputation with receiving mail servers. Ken manages warmup through multiple providers simultaneously. Warmup ramps follow hard-coded rules (never more than +20% per day, pause when bounce rate exceeds 5%). An inbox must reach a configured minimum warmup age before Ken will include it in a campaign's sender pool.
In the app: Email Delivery - Warmup.
Workspace
A tenant account in Ken AI. Each workspace has its own campaigns, contacts, inboxes, domains, and billing. Users can belong to multiple workspaces and switch between them from the top navigation.
Troubleshooting & debugging
Symptom: A campaign shows 0 sent emails despite being Active. Likely cause: no sender config attached, or all attached inboxes have a warmup age below the minimum gate. What to check: Campaigns - Actions - Senders. Verify at least one sender config is attached and that the inboxes in it show a delivery score and a warmup status of Active.
Symptom: A contact is not moving through enrichment. Likely cause: the email address was marked Invalid by a prior verification run. Invalid is permanent. What to check: open the contact record and inspect the email_validity field. If it is 4 (Invalid), the contact will never enrich and cannot receive mail. Remove from the campaign or upload with a different address.
Symptom: Deliverability score dropped sharply overnight. Likely cause: a DNS health change (DKIM key rotation, DMARC misconfiguration) or a new DNSBL listing. What to check: Email Delivery - Inboxes - select the affected inbox - DNS tab. Look for a failed SPF, DKIM, or DMARC check. Also check the Blacklists tab for any new listings.
Symptom: Analytics show 0 clicks but the email clearly had a tracked link. Likely cause: the click count on the Analytics page reads from Short.io tracking links. If the link was not set up as a Short.io redirect, Ken records no clicks. What to check: open the campaign's sequence and confirm tracked links are rendered as Short.io short links (they appear in the preview as short URLs). If they are not, re-save the sequence step with tracking enabled.
Symptom: Placement test shows Spam for a healthy-looking inbox. Likely cause: domain reputation, content flags, or a missing DMARC record. Placement tests send real campaign copy, so copy with spam trigger words will fail. What to check: Email Delivery - Domains - check the DNS badge column for the affected sending domain. Confirm SPF, DKIM, DMARC, and MX are all green. If DNS is clean, review the email copy in the placement test for spam-trigger phrases.
Symptom: A contact received the same email twice. Likely cause: the contact was imported into two different segments or two different campaigns without the dedup window enabled. What to check: Campaigns - Settings - Recently Emailed window. If the window is off, contacts can receive mail from multiple overlapping campaigns. Enable the window or check for duplicate segment membership.
Symptom: AI flow is not running for some contacts. Likely cause: the per-stage AI contact cap was reached, or the AI workflow is paused. What to check: Campaigns - AI tab - check the flow's contact cap setting. Also confirm the workflow is not in a Paused state. Each flow has an independent budget limit.
FAQ
Q: What is the difference between a segment and an audience? An audience defines who you want to target (search filters, CSV upload, KenSearch import). A segment is a named subdivision inside a campaign with its own prompts and sequence. An audience feeds into a segment; a campaign can have multiple segments with different audiences.
Q: When does warmup need to be complete before a send? Ken enforces a minimum warmup-age gate. By default, an inbox must have been warming for at least 14 days before it is eligible to send. The exact threshold is configurable per workspace.
Q: What does "Sticky Invalid" mean? Once any verification provider returns an Invalid verdict for an email address, Ken marks that address Invalid permanently. The verdict persists across all future campaigns and cannot be overridden manually. This prevents wasting enrichment credits and sending to bad addresses that have already been confirmed as undeliverable.
Q: Can I change the AI model mid-campaign? Yes. Model settings are per-flow and per-campaign. Changing the model after launch applies to contacts that have not yet run through that flow. Contacts that have already been personalized are not re-processed unless you manually re-run the flow.
Q: What is the difference between a Batch and a Trigger campaign? A Batch campaign sends emails to all contacts on a schedule - it blasts the full list over time. A Trigger campaign sends one personalized email per contact the moment a specific intent signal fires. Trigger campaigns require a signal rule and a sender config; they do not send until a signal event occurs.
Q: Why does the sequence editor show both Email 1 and Email 1a? The "a", "b", "c" suffix indicates an A/B variant. Email 1 is the base version; Email 1a is the first alternate variant for the same step. Each variant is sent to a configured percentage of contacts in that segment.