01
Signal
Spots the moment that matters.
Trends, mentions, intent scores on one side. Stalled deals, ICP fit, unread threads on the other. Wyse watches both.
PostWyseIntent · 92mention
OpsWyseStalled · 14ddeal
intent scoring · pipeline priorities
Meet Wyse
Always on · v2.1One AI. Every surface we ship.
Wyse is the always-on agent inside both PostWyse and OpsWyse. Over 20 distinct capabilities, one brand, one voice, tuned per surface. Drafts the move, ranks the next one, hands it to your team with context.
Watches 24/7Drafts in your voiceHands off to your teamPer-workspace isolationBYO API key
The Wyse loop
From signal to closed loop.
Wyse doesn't sit in a chat box. It runs the same three beats inside every surface. Content on one side, revenue on the other, the same loop each time.
Runs per signal · avg 1.8sHuman approves every send100% provenance
02
Draft
Writes in your voice.
Brand-kit aware. Word-limit disciplined. 0.8s per draft. Every line tagged with provenance before it leaves.
PostWyseBlog openerSEO · 94
OpsWyseReply98 / 120 w
temp 0.3 – 0.4 · brand-kit aware
03
Handoff
Lands with your team.
One click pushes the lead to Leorix with full context, or creates a task from the next-action. Nothing sends without a human.
drafted_bywyse9:24 AM
reviewed_bymia.k9:25 AM
sent_byawaiting–
Leorix push · create task
Inside PostWyse
Fifteen content moves. One agent.
Wyse inside PostWyse spans the full content loop, from detecting a lead on X, to the AI Content Planner nothing else in marketing has, to Wyse Embed knowing your whole business, to Wyse Campaigns shipping email like Mailchimp without leaving Wysera.
Lead detection · Marquee
Intent scoring, 0 to 100.
Every detected lead (from mentions, inbox DMs, and form submissions) gets a 0–100 intent score from engagement signals. Company, role, platform source, original message, the signals that contributed.
Acme92
Vanta87
Mercury76
Linear63
Raycast44
Filter 70+ · 80+ · 90+Hide handledintent · live feed
Brand intel · Marquee
Brand kit from a URL.
Crawl any site. Wyse extracts the full kit: palette (hex), fonts, voice, archetype, competitors, pricing. Grok-3 with a 100s timeout.
Palette
#0a0a0b
#c2410c
#f97316
#fb923c
#fed7aa
Type
Fraunces
Inter · body
Archetype
Friendly challenger
Voice: direct, warm
POST /api/ai/brand · Grok-3
AI Content Planner · FlagshipNowhere else
Plan the whole month, in seconds.
An AI-native planner you won't find anywhere else in marketing. Wyse reads your brand voice, search demand, trending topics and past performance, then proposes a full monthly map and auto-schedules every slot across every channel.
May 2026 · Proposed mapGenerated 0.9s
Launch daySocialBlog
30-day map · auto-schedule · learns what lands
Wyse Embed · Site chatbot
A chatbot that knows your business.
Drop one embed on any landing page. Wyse becomes your AI support and sales rep, trained on your content, brand voice, pricing and FAQs. Leads flow straight into OpsWyse.
WyseDo you integrate with HubSpot?
Yep. We ship a Zapier bridge and a native HubSpot import on Pro. Want me to pull the setup doc for you?
One embed · full brand knowledge · leads hand off to CRM
Wyse Campaigns · FlagshipReplaces Mailchimp
Email campaigns, in your voice.
Drag, drop, ship. AI drafts the subject in your brand voice, Wyse picks the send time, A/B tests what headline lands, and drip flows you build in two clicks. Replaces Mailchimp, Brevo, and Mailerlite without leaving Wysera.
Drag-drop builderA/B testingDrip automationsTemplatesAudience segmentsOpen + click trackingSend-time AICSV + Mailchimp import
Brand-aware drafting · CRM-aware audiences · runs inside Wysera
Campaign · Q2 launchSent · 14:02
Subject
The 3 hours you waste every Monday
A · 64% openB · 47%
Open
64%
Click
18%
Reply
4%
Wyse draftedsend time · Tue 9:15 AM
Reply drafting
Replies in your voice
One click drafts a contextual reply using your stored brand guidelines. Editable inline before sending.
Grok-3-fast default
CRM handoff
Push to Leorix, in one click
Lead + reply + intent score + signal breakdown + source platform, bundled and pushed to the CRM in a single call.
POST /api/crm/leorix/push
Content generation
Titles, captions, CTAs
Per platform, per content type, enriched with Semrush search volume and difficulty on every keyword.
POST /api/ai/generate
Ideation
Topics, campaigns, keywords
Wyse returns a JSON list of content ideas for any topic or keyword set. Each includes title, description, type, and tags.
titledescriptiontypetags
Competitor intel
Analyze, or discover four
Analyze
URL in. Company, industry, target keywords out.
Discover
4 competitors with strengths, G2 reviews, pricing.
POST /api/ai/competitors
Keyword research
Suggest + enrich
Wyse suggests, Semrush enriches each with volume, CPC, and difficulty so you don't guess what ranks.
Vol12.4k
CPC$4.20
KD38
AI visibility
Rank inside AI answers.
Tracks where your content surfaces in each model's answers, not just classic Google SERP. AIO and GEO coverage out of the box.
ChatGPTClaudePerplexity
Publishing
Six networks, one queue
publishWorker ships scheduled posts with native schedules per channel. No copy-pasting across six apps.
LinkedInXFacebookInstagramThreadsMedium
Listening
Mentions + trends, live
listeningWorker pulls brand mentions and trending topics. The raw feed the Agent reads to score intent in production.
Worker: listeningWorker
Metrics
Engagement, synced
metricsWorker syncs engagement from every connected platform. Analytics isn't a tab you switch to. It's ambient context for the Agent.
Worker: metricsWorker
PostWyseWhere it lives
5 surfaces · live- Wyse Agent tabTab
Lead feed, filters, triage, draft, handoff to CRM
/dashboard?tab=wyse-agent - Sidebar pinPin
One-click entry to the Agent from anywhere
Social Hub · NEW badge - Unified InboxInbox
Invoke the Agent from any DM or comment
Conversation threads - Events CRMModule
Score and hand off captured contacts to Leorix
Captured contacts view - Content CalendarCalendar
Draft posts in bulk, queue every channel
Scheduling surface
Inside OpsWyse
Fifteen CRM capabilities. One outbound engine.
Wyse inside OpsWyse is a cross-surface copilot. It ranks the pipeline, drafts the call prep, and runs the new Wyse Outbound engine, finding leads from your ICP, our 1M-record database, your Apollo key or our Chrome extension, then auto-pitching each one based on personality.
Pipeline priorities · Marquee
Ranks every deal. Flags the stales.
Up to 50 open deals in one call. High / medium / low urgency, with a dedicated bucket for anything 14+ days silent. Cached 15 minutes on the dashboard.
highAcme Inc.$48kReply drafted · awaiting send
highVanta$120k2 days silent · call prep ready
mediumMercury$36kProposal sent · follow Thu
lowLinear$22kNewly added · research phase
staleRaycast$80kStalled 18 days · no activity
/api/ai-pipeline-prioritiestemp · 0.2max · 50 records
Sales brief · Marquee
Call prep, five sections.
Terse, fixed-shape brief in the record drawer before every call. No filler, no reshuffling.
- 01
Where we left off
Pricing for the team tier
- 02
Their stated interest
Scaling automation across ops
- 03
3 questions
Renewal · budget owner · competitors
- 04
Goal
30-min demo → pilot offer
- 05
Opening angle
ROI deck shared Mar 12
/api/ai-summarize · call_prep
Wyse Outbound · FlagshipReplaces Apollo + Outreach
Find them. Pitch them. All on autopilot.
Wyse reads your company, infers your ICP, then pulls a matched list from our built-in million-record database, your Apollo key, or our Chrome extension scraping any page. Every lead gets a personality read and an auto-drafted pitch in your voice. You approve. Wyse sends.
ICP auto-detect1M+ lead databaseApollo BYO keyChrome scraperPersonality readAuto-pitch in your voice
Pulled from your ICP · enriched · pitched · queued
ICP · B2B SaaS, 50-500, RevOps12 fresh
- SP
Sara Patel
VP RevOps · Stride
Fit 96DB - MR
Marc Rivera
Head of Sales Ops · Lattice
Fit 88Apollo - JC
Jenna Cho
Director Ops · Atlas Studio
Fit 74Chrome
Wyse drafted for Sara
“Sara, I noticed Stride moved RevOps in-house last quarter. Two specifics on how we cut deal-cycle 22% below.”
Record summary
Web-enriched record brief
Narrative summary in the record drawer. Markdown-lite, 800 tokens max, enriched with the web.
/api/ai-summarize · summary
Thread summary
Email thread in five
Current state · Their asks · Your commitments
Open threads · Suggested next move
/api/ai-summarize · thread_summary
Friday recap
Weekly narrative card
MovedStalledWinsRisksMonday
/api/ai-summarize · weekly_recap
Reply drafting
Warm, direct, under 120 words.
In-thread. First-name greeting. No fabricated facts. Your brand voice, drafted in 0.8s.
98 / 120
First-touch
Cold outreach, <110 w
One clear next step. Zero buzzwords. Pulls the record title as the intended subject.
/api/ai-draft-reply · new
Next action
One move, today
≤ 18 words. Plus the reason, ≤ 16.
suggestionreason
Triage mode · Dark accent
Keyboard-driven inbox zero.
One card at a time. The queue is Wyse picks ∪ stalled ∪ due-soon, minus anything in closed stages. Keys to act, approve, or dismiss.
Queue composition
1Wyse picks (intent + fit)
2Stalled · 14+ days silent
3Due-soon · commitments
Keys
JnextKprevEapproveXdismissRreply↵open
NL → grid
Plain English, table actions
> Add 5 deals for Q2 medical sector
add_rowupdate_rowdelete_rowsortbulk_updateget_row_info
Voice capture
Speech → fields
/api/ai-voice · temp 0.1
Filter parsing
NL → filters
industrysizetitlecitycountrystate
Row generation
Stream ≤ 50 rows
SSE · temp 0.8
Intent routing
6 intents, deterministic
pipeline_mgmtdata_opspipeline_querymodule_setupgeneral_querypipeline_cfg
OpsWyseWhere it lives
10 surfaces · live- WyseSummaryDrawerDrawer
Summary and call-prep brief with tabs
Record detail · right side · 440px - WyseNextActionPanelPanel
One concrete next action + create-task
Top of record detail - EmailThreadSummaryInline
Inline 5-section thread brief
Every email thread header - InlineReplyComposeComposer
Draft reply with Wyse, editable before send
Email reply surface - InlineComposeNewComposer
First-touch outreach drafting
New email compose - TriageReplyDrawerDrawer
Reply overlay with keyboard shortcuts
Full-screen Triage Mode - TodaysPrioritiesCardCard
Daily deal ranker with template picker
Home dashboard - FridayRecapCardCard
Weekly recap narrative card
Dashboard · Friday afternoon - WysePipelinePanelPanel
Focus today + Stalling (14+ days)
Pipeline hero area - AIGridChatbotChat
NL → grid actions, SSE streaming
Table surfaces
Under the hood
Opinionated about how Wyse thinks.
Every AI call has a model, a temperature, and an output shape chosen for the job. No generic chatbot knobs. You can also bring your own Claude or Grok key and override the defaults per org.
Model mix
- OpsWyse engineGPT-4o-mini
All 10 AI routes. Voice set per-route in system prompts, not model choice.
- PostWyse defaultGrok-3-fast
Speed + cost-optimized for content routes. Logged per org and user.
- Brand extractionGrok-3
Higher token budget for full brand-kit JSON. 100s timeout.
- Quality fallbackClaude
Org-level Claude key supersedes env. OpenAI supported via the same BYO slot.
Temperature discipline
- 0.1Deterministic
Intent routing, voice extraction, filter parsing. No variance allowed.
- 0.2Structured
Grid actions, pipeline priorities. Shape-constrained output.
- 0.3Narrative
Summaries and single next-action suggestions. Controlled prose.
- 0.4Email drafts
Replies and first-touch outreach. Slight variation for warmth.
- 0.8Diversity
Template row generation. Forces realistic spread across rows.
Output formats
- Strict JSON · priorities, next action, intent, voice, filters, grid
- Markdown-lite · summaries and reply HTML
- SSE streams · grid chat, row generation
Provenance
Every action writes drafted_by, reviewed_by, and sent_by with timestamps. Audit any customer-facing draft end-to-end.
Workers
publishWorker · 6 networks. metricsWorker · engagement sync. listeningWorker · mentions + trends. engagementSyncWorker · unified inbox.
Logged per org & user · src/lib/logAiUsage.tsProvider priority · Org Claude → Env Claude → Org Grok → Env Grok
Why trust Wyse
Built so you can actually use it.
Three commitments we wrote into the system, not the marketing.
The Wyse Trust Pledge
- 01
We won't
Train public models on your data.
Wyse learns per-workspace to hold your brand voice. Nothing trains public models. Nothing leaks between workspaces. Your workspace sees only your workspace.
No cross-workspaceGDPR-safeNo cookiesPer-org encryption - 02
We always
Stamp every draft with provenance.
drafted_by · reviewed_by · sent_by with timestamps. Audit every AI touch end to end. If Wyse wrote a line, you can see it. If you edited it, you can see that too.
drafted_byreviewed_bysent_bytimestamp-logged - 03
We never
Send without a human in the loop.
Wyse drafts, ranks, and suggests. A human on your team approves every customer-facing action. Autopilot exists, but it's opt-in, never default.
Human approvesOpt-in autopilotNo auto-sendReviewer state required
Signed, Wysera
Effective · 2026-04-23 · v1.0
FAQ
Questions about Wyse.
The things people ask before they hand an AI access to their inbox and editorial calendar.
What is Wyse?
Wyse is the AI copilot built by Wysera. It runs inside both PostWyse (content, SEO, social, brand intelligence) and OpsWyse (CRM summaries, replies, priorities, Triage Mode). One brand, one voice, over 20 distinct capabilities tuned per surface.
What AI models does Wyse use?
OpsWyse runs on OpenAI GPT-4o-mini across all 10 API routes, with temperature tuned per task (0.1 for routing, 0.3 for narrative, 0.4 for email drafting). PostWyse runs primarily on Grok-3-fast for speed and Grok-3 for heavier work like brand extraction, with Claude as a quality fallback. You can bring your own Claude, Grok, or OpenAI key at the org level to override defaults.
Does Wyse train on my data?
No. Wyse learns per-workspace to stay in your brand voice, but nothing trains public models, nothing leaks across workspaces, and every AI draft carries its provenance (drafted_by, reviewed_by, sent_by with timestamps) so you can audit exactly what shipped and why.
Can I use Wyse without PostWyse or OpsWyse?
Not today. Wyse is the AI surface inside the two products. That tight integration is what lets it act on real signals (content trends, trending mentions, deal activity, inbox threads) rather than generic chat.
How is Wyse different from ChatGPT or Claude?
General assistants give you a chat box. Wyse lives inside the specific surfaces where you work: the editorial calendar, the pipeline board, the email thread, the record detail drawer. It watches the moments that matter, drafts in your voice, and hands the next move to your team with full context.
Does Wyse take actions without approval?
Never on customer-facing work. Wyse drafts replies, posts, and moves with word-count discipline (replies under 120 words, cold outreach under 110 words) but a human on your team approves before anything sends. Autopilot is opt-in, never default.
Can I bring my own API key?
Yes. Provider priority is: org-level Claude key → env Claude key → org-level Grok key → env Grok key → OpenAI (via the same BYO slot). Every call is logged per org and user for usage analytics and future billing.
Are Wyse features streaming or batch?
Both. Narrative summaries and structured JSON outputs (priorities, next actions, classification) are batched. The table chat (AIGridChatbot) and row-generation use Server-Sent Events so you see the answer type. Email drafts are batched since you want to review the full draft.
How does Triage Mode work?
Triage Mode is a keyboard-driven inbox-zero flow over Wyse-flagged records. The queue is Wyse picks unioned with stalled deals and due-soon records, minus anything in closed stages. One card at a time, keyboard shortcuts to act, approve or dismiss.
Is Wyse customizable per workspace?
Yes. Each Wysera workspace gets its own tuning: brand voice, ICP fit, preferred channels, approval rules. Wyse in one workspace never sees or learns from another. Brand kits extracted by /api/ai/brand are saved per org so every generation stays on-brand.
Meet Wyse inside the work
Wyse doesn't live in a chat box. It lives in yours.
Spin up a Wysera workspace and Wyse is already inside the editorial calendar, the inbox, the pipeline, waiting for the next signal.