The internal home for every human role and the AI agents they will train, direct, and supervise. Skills are the building blocks; agents are what we're building on top of them — persistent workers with specific jobs, inputs, outputs, and cadences. Each human is the direct report for their agent team.
Keeps Nick's priorities, open loops, and strategic initiatives organized across the company.
Build planComposes TaskTracker push + Fathom task extraction + calendar context. Daily digest at 6am summarizing open loops, overdue commitments, and the day's gates. Human gate: Nick reviews and approves before any task is auto-assigned to the team.
Deep research for offer design, market entry decisions, and competitive positioning.
Build planComposes /market-scan + /local-seo-competitors + WebSearch + Clawton. Triggered on demand when Nick asks a strategic question. Output is a briefing doc with sources, not recommendations — Nick stays the decider.
Pipeline health, forecast, win/loss analysis, and source attribution across GHL + prospector + inbound.
Build planReads GHL MCP + Fathom call history + Local Prospector pushes. Weekly Monday forecast report + monthly attribution analysis. Flags stalled deals and proposes specific re-engagement moves.
Coordinates content across BCT + Tekton site + social + lead magnets. Maintains brand voice guardrails.
Build planComposes /bct-draft + /youtube-strategy + humanizer + social calendar. Monthly: produces the content calendar. Weekly: flags gaps. Anne executes; Nick approves. Guardrail: no em dashes, no "one per city" claims, no lead-tool bashing.
Coordinates the agent platform itself — priorities, builds, deprecations. Eventually delegates to specialized agents.
Build planFour-tier permission model. Review responses first, expand task by task. Anti-rubber-stamping rules. Becomes the spine that coordinates every other agent once it ships. This is the "platform operator" role being built before Nick hires one.
Kyle owns the client relationship and orchestrates fulfillment. Alessandra leads technical execution; Edward, Jakob, and Anne own their specialist lanes. Each human trains their own agent roster — agents compose the skills we have today into persistent workers with defined inputs, outputs, and cadences.
Daily portfolio health scoring — who's green, yellow, red, and why.
Build planReads TaskTracker activity + Local Dominator rank trends + GA4 traffic + GSC alerts + review sentiment. Daily 8am brief to Kyle flagging at-risk accounts before client notices. Triggers: ranking drop >10% of keywords, missed delivery, no client response >2 weeks.
Drafts win updates, check-ins, recap emails, CSAT prompts in each client's preferred tone.
Build planComposes /fathom-tasks call context + CSM email template library + client-specific voice notes. Draft → Kyle edits → send. Human gate: nothing goes out without Kyle's approval. Minimum 2 win updates per client per month auto-queued.
Compiles the monthly deliverable for every client — rankings, traffic, content, backlinks, technical wins.
Build planComposes /seo-monthly + /seo-report + /seo-data. Runs 1st-3rd of each month, drops drafts in Kyle's queue. Kyle personalizes and QAs. Kills the report-building bottleneck that blocks the 5th-of-month SLA.
Builds a prep doc 24 hours before every scheduled client call — talking points, data, open items.
Build planReads last 30-day Local Dominator + GSC + published content + open TT tickets + last call's Fathom notes. Auto-creates the brief as a TaskTracker note tagged to the call. Kyle reviews and edits before the call.
Runs the 90/60/30 day renewal path — health check, value recap, pricing scenarios.
Build planTriggered 90 days before contract end. Produces value-delivered recap + ROI math + risk flags + renewal script tailored to client sentiment. At-risk renewals get escalated to Nick automatically.
Weekly AEO + Google visibility scan across all clients. Running on Clawton today.
Build planEvery Monday 7am ET on VPS cron. Regressions hit Kyle's Telegram. Next evolution: auto-open TaskTracker tickets on regression with recommended fix wired to Alessandra.
Already deployed via Clawton — watches reviews every 4 hours, routes to Edward.
Build planEdward's lane today. Kyle sees rollup in portfolio health. Next evolution: Kyle gets notified on critical reviews (1-3 star) so CSM can coordinate client outreach before Edward posts response.
Runs every monthly check across the entire roster and produces phase gate decisions per client.
Build planComposes /seo-batch + /seo-monthly + /seo-crawl + /seo-data + /seo-gap-analysis. Runs 1st of month. Output: per-client phase (Core 30 / Topical / Geographic / Technical) + prioritized tasks. Alessandra reviews, approves, pushes to TaskTracker via existing script.
Drafts service pages, location pages, and topical articles end-to-end — ready to publish, not ready to rewrite.
Build planComposes /seo-write-page + /local-seo-copywriting + /local-seo-topical-content + humanizer. Triggered from the Pipeline Conductor's task list. Guardrails: contractor language (not SEO jargon), no em dashes, no "one per city," hardscape metaphors when Nick's voice fits.
Continuous scan for schema gaps, Core Web Vitals regressions, indexing issues, and mobile errors.
Build planComposes /seo-crawl + /local-seo-schema + GSC data + Lighthouse. Weekly scan per client. Output: ranked fix list with deploy-ready schema JSON-LD blocks. Alessandra deploys or delegates to Jakob.
Converts heatmap scans into monthly geo-content plans — neighborhood targets, landmark mentions, pocket fixes.
Build planComposes /local-seo-heatmap + /seo-geo-strategy + /local-seo-location-pages. Monthly per client. Classifies pocket vs wide issues, picks 2-4 monthly actions. Quality gate blocks thin location pages before they ship.
Monitors and improves visibility in ChatGPT, Perplexity, Google AI Overviews, Gemini.
Build planComposes /ai-search-optimization + /local-seo-ai-visibility + weekly AI query tests. Flags when client drops out of AI recommendations. Produces entity optimization fixes + content framework updates.
Internal linking plans, citation cleanup, and local sponsorship/chamber opportunities per client.
Build planComposes /seo-links + /local-seo-citations. Monthly internal link plan with anchor text. One-time chamber/sponsorship research (1-2 per client per year). Outputs queue into Edward's BrightLocal workflow.
Drafts all 4-5 star review responses, sends to Edward in TaskTracker chat for approval.
Build planAlready deployed on Clawton VPS as of April 2026. Every 4 hours: pulls new reviews across all clients, drafts personalized responses in client voice, asks Edward a verification question (prevents rubber-stamping). Edward answers + approves/edits/rejects. Approved responses created as "Post to GBP" tasks. Next evolution: auto-post after approval.
Builds monthly GBP posting calendar and drafts every post in client voice with images attached.
Build planReads posting-calendar.md + client brand voice + service list + seasonal context. Produces 4+ posts per client per month with CTA, image, keywords, 300-500 char count. Edward reviews in Search Atlas, edits, schedules. Guardrails from the SOP baked in (no placeholder text, no duplicates, mix of themes).
Native publishing to GBP without Search Atlas dependency — our own product for clients.
Build planGoogle API application submitted 2026-04-16 via info@tektongrowth.com. Cloud project artful-sky-473915-j9. Build kicks off on approval. Eventually ingests the GBP Post Author Agent's drafts and posts directly.
Continuous NAP consistency scan, duplicate detection, cleanup queue.
Build planComposes /local-seo-citations + BrightLocal API. Weekly scan per client. Output: ranked cleanup list (duplicates first, then inconsistencies, then missing high-priority). Edward works the queue in BrightLocal.
Sources fresh GBP photos monthly from Instagram, website, reviews. Tags by category.
Build planPulls from client Instagram + website gallery + GHL uploads. Rotates 2-3 fresh photos per client per month. Categorizes: exterior, interior, staff, products, before/after, seasonal. Edward approves + uploads.
Monthly profile audit — categories, services, attributes, hours, photos, posts, Q&A.
Build planComposes /local-seo-gbp + /local-seo-landing-page-audit + /gbp-rollout. Monthly deep audit per client. Flags drift: hours changed, services mismatched between GBP and site, category recommendations missed.
Runs intake calls (with Kyle + client), extracts requirements, produces a locked build brief.
Build planComposes /fathom-tasks call parser + brief template + reference site analyzer. Takes call recording + client inspiration URLs, outputs: brand tokens, feature list, content outline, imagery gaps, risks. Jakob reviews, confirms with client, ships as the build spec.
Multi-agent orchestrator that routes through intake, config, copy, images, audit, deploy. First shipped CM Construction April 2026.
Build planRunning today. 4 subagents (copy, images, seo-audit, deploy) + astro-site-generator engine with 12 templates. Next evolutions: refine-template subagent integration + Jakob-in-the-loop checkpoints between phases instead of end-of-run review.
Blocks deploy when anything in the QA checklist fails — link rot, form failures, broken tracking, heavy images.
Build planExtends current website-seo-audit subagent. Automated link validation + Lighthouse + mobile rendering + form submission tests. Produces a pass/fail report with specific blockers. Jakob gets a green light or a fix list — deploy impossible until green.
Handles GA4 + GTM + GSC + OTTO install and verification across every platform.
Build planComposes /gsc-verify + GTM template + platform-specific install scripts. Verifies live events fire correctly. Runs in the 48-hour pre-launch window automatically. Covers WP, Squarespace, GHL, and Astro.
Takes a reference image + current build, iterates until 1:1 visual match. Template-level or section-level.
Build planBased on /refine-template skill. Screenshot compare loop: render → compare → adjust → repeat until pixel match. Respects layout-only vs colors boundary (brand tokens flow from site-config, never hardcoded). Human gate: Jakob confirms after every iteration pass.
Monthly security + plugin updates across every managed WordPress site, with backup verification.
Build planReads plugin inventory per client site. Monthly: triggers updates in a test environment, runs QA checks, promotes to prod if green. Backup verified before + after. Rollback automatic on regression. Critical security patches push in 48 hours with Jakob notified.
Daily triage of all new LSA leads across every client account by 10am.
Build planPulls new leads from Google LSA dashboards. Flags suspicious patterns, categorizes by service match, prioritizes by quality signals (lead details, contact completeness). Anne reviews the triaged queue and actions.
Monitors client spend against max budget, toggles ads ON/OFF at threshold, never lowers budget.
Build planReads LSA spend daily. Toggles OFF when client hits max budget for the month. Toggles ON first day of new month. Logs every action. Guardrail from SOP: never lower the budget (affects ad rank) — spend control is toggle-only.
Runs the bi-weekly lead quality form to clients; uses feedback to retune targeting.
Build planSends GHL Survey form every 14 days. Aggregates responses, maps "great leads" to common patterns, proposes targeting adjustments in LSA dashboard. Anne approves the adjustments before they push.
Monthly performance report assembler — per-client LSA deliverable for Kyle.
Build planComposes /pdf-report-design + LSA dashboard data + client feedback form results. Runs 15th of each month. Anne QAs and ships to Kyle by the 20th. Kills the manual spreadsheet bottleneck.
Triages inbound internal tasks and routes to the right team member in TaskTracker.
Build planReads Tekton Growth Internal TT project. Auto-tags: client-specific vs platform vs marketing, urgency, likely owner. Proposes assignment; Anne confirms in chat. Uses direct Postgres pattern for subtask creation.
Handles the post-production pipeline for Blue Collar Techy videos — captions, thumbnails, descriptions, Shorts.
Build planComposes /youtube-seo + /youtube-shorts + /yt-thumb + /youtube-script-polish + humanizer. Triggers on new video upload. Output: full metadata package + 3 Shorts concepts + pinned comment. Anne ships.
Executes the Tekton + BCT social content calendar — drafts posts, schedules across platforms.
Build planReads monthly content calendar from Marketing Director Agent. Generates platform-specific variants (IG, FB, X, LinkedIn). Schedules via platform APIs. Anne reviews weekly. Guardrails: short punchy statements only, no story-style narrative.
Generates FB ad variants and LSA ad copy per client; helps diagnose underperformers.
Build planReads client brand voice + past high performers + offer details. Outputs 5-10 copy variants per test cycle. Anne picks, launches, tracks performance. Agent learns from what wins.
Drafts community responses across audience touchpoints — YouTube comments, FB groups, DMs.
Build planReads Nick's voice profile + community context. Drafts on-brand responses. Anne approves and posts. Flags high-value engagements (potential leads, partnership signals) for Nick.
Nick runs sales strategy and closes. Rico is developing under direct mentorship — outreach is his first lane, with a path toward full AE responsibilities as skills compound.
Automated lead gen — scrapes Google Maps + Belgard, filters by size, checks SERP, pushes to GHL with NEPQ script attached.
Build planRunning today. Next evolution: continuous mode — runs weekly per market, deduplicates against existing pipeline, prioritizes by fit score (size × SEO gap × competitive weakness).
Personalized cold outreach per prospect — not templated spam. Uses real intel from Local Prospector reports.
Build planReads prospect's SERP weakness + competitive position + services list. Drafts opener that references a specific observable gap. Rico edits + sends via GHL. Agent learns from reply rates which angles land.
Builds a prospect brief before every discovery call — rankings, gaps, competitive pressure, talking points.
Build planTriggered when GHL moves a prospect to "Call Scheduled." Composes /ai-search-lead-audit + SERP analysis + past message history. Brief drops in Rico's Slack/TT 24 hours before the call.
Scope-based ROI proposal — search volume × local population × call rate × close rate math.
Build planRunning today. Next evolution: auto-trigger after every discovery call that ends in "send me a proposal," pre-filled with scope from call notes. Rico reviews, personalizes, sends.
Free AI Search audit as a cold-outreach door-opener — branded PDF with prospect's gaps.
Build planRunning today. Runs on demand or auto-batch. Pairs with Outreach Drafter Agent: first-touch sends the audit, second-touch opens the conversation.
Reviews Rico's Fathom call recordings against the NEPQ framework rubric and gives structured feedback.
Build planComposes Fathom transcript parser + NEPQ rubric (problem awareness, consequence, commitment, etc.). After every recorded call, delivers a scored breakdown to Rico + Nick. Highlights exact moments to study. Nick still does live mentoring; agent handles the pattern work.
NEPQ-style 3-email sequence for cold leads who took a call but didn't sign.
Build planRunning today. Next evolution: auto-trigger from GHL pipeline stage change (Call Completed → No Signup for 7+ days). Rico reviews the drafts before send.
Blue Collar Techy is Nick's personal brand and Tekton's top-of-funnel channel. Nick is on camera; Anne produces and repurposes. Tekton Growth is its own media property being rebuilt from GHL to custom Astro.
Drafts and revises BCT blog posts with current-events research and Clawton voice pass.
Build planRunning today. Guardrails: contractor language, no em dashes, no spreadsheet recommendations (BCT positioning is practical tech), mortgage rates + consumer sentiment cited when writing about market conditions.
Full script pipeline — topic → research → hook → script → polish in Nick's voice.
Build planComposes /youtube-strategy + /youtube-script + /youtube-script-polish + humanizer. Voice profile pulled from Fathom transcripts. Nick picks from hook variants, edits, records.
Monthly YouTube audit — hook quality, retention signals, pillar gaps, Shorts balance.
Build planComposes /youtube-audit + YouTube Analytics API. Monthly report with specific actions ranked by impact. Nick and Anne action the top 3.
Autonomous TikTok slideshow creator for Tekton Growth and BCT brands.
Build planIngests BCT videos + blog posts + market-scan data. Generates 1-2 slideshow posts per day per brand. Nick reviews batch weekly. Fully autonomous execution once voice is dialed in.
The shared infrastructure every agent runs on top of. Nick owns this layer today. Eventually spins out as a dedicated Platform Operator role.