AI Agents We Implement

The AI agents we build with, and how we pick yours

We don’t sell one tool. We pick the agent that fits the job, connect it to your apps through Composio, and build it inside your own accounts. Here is what each agent does, what the build looks like, and the demo you will see.

Alyzza from our team reviewing an AI agent dashboard on a laptop

Four agents, one orchestrator, built in your accounts.

How It Fits Together

Three layers, all in your accounts

The agent

Thinks, writes and decides

The rails

Runs fixed steps on time, every time

The orchestrator

Composio gives any agent secure access to your CRM, inbox, calendar, Slack, Sheets, WhatsApp and hundreds of other apps. Logins are managed in one place and can be revoked anytime, and the same connections carry over if you switch agents later.

Pick In 30 Seconds

Which agent fits the job?

Most builds use two: one agent for the thinking, and Make or n8n for the rails.

Claude CodeGrok BotPerplexityMake / n8n
Best atJudgment-heavy work: writing, CRM notes, proposals, reports, codeAlways-on tasks across your apps, X signals, low-cost voiceDeep research and reports, cross-checked across modelsFixed, high-volume pipelines that must run the same way every time
Where it runsYour Claude account; scheduled tasks run in the cloudPersistent agents, each on its own cloud computerWeb, Mac app, Slack and TeamsYour Make account, or n8n cloud or self-hosted
Day to dayYou approve drafts in SlackYou assign tasks in plain EnglishYour team asks; Projects keep the contextNothing to do; you get an alert if a step fails
Running costAbout $20 to $200 a month per userSeat plans; voice about $0.08 a minuteMax about $200 a monthLow monthly plans
Demo you'll seeLead in, reply drafted, CRM updated, approval in SlackCall the voice agent, or watch it work three appsA sourced market or account report in minutesTest form to WhatsApp reply, booking and CRM in under a minute

Prices are approximate as of November 2026 and change often. We confirm current pricing on your strategy call.

01

The brain for judgment and writing

Claude Code

Anthropic's agent that works inside your files and apps. We use it for any job that needs thinking and good writing, not just moving data from A to B.

How it helps you

  • Takes over repeat knowledge work: research, drafting, reporting and follow-up, done in your voice.
  • Learns your business from a written playbook and skills, so the output gets sharper every week.
  • Runs on a schedule in the cloud, so the work is done before you open your laptop.

Use cases

  • Outbound: researches each target account, writes the first line and logs it in your CRM
  • Sales: turns a call recording into a branded proposal deck the same day
  • Content: turns one voice note into a week of posts in your voice
  • Reporting: pulls CRM, ads and Stripe numbers into a Monday report in Slack
  • Websites: builds and updates your site from plain-English requests
  • Finance: models cash runway and margins from your spreadsheets

If you choose it, here is the build

  1. Week 1Access and context

    Your Claude Team or Max account, your apps connected through Composio, and the playbook file that teaches it your offer, voice and rules.

  2. Week 2Build the skills

    One skill per job, such as qualify a lead or draft a proposal, tested on your real data.

  3. Week 3Schedule and approve

    Scheduled tasks run on their own and post drafts to Slack. Nothing goes out without a human yes during the pilot.

  4. Week 4Handover

    Runbook, a 60-minute training for your team and a 30-day check-in.

The demo you will see

A new lead lands. The agent researches the company, drafts a reply, updates the CRM and asks you to approve it in Slack, live on the call.

Running cost

Included in Claude plans, from about $20 to $200 a month per user. Team seats suit small teams.

What you own

The playbook, skills and scheduled tasks live in your Claude account and your files.

Watch-outs
  • Heavy daily jobs use more of your plan, so busy builds usually need Max or Team seats.
  • It sounds like you after about a week of feedback, so we keep an approval step at first.
02

The always-on teammate

Grok Bot

xAI's persistent agents. Each one gets its own cloud computer and signs into your tools, so it keeps working through the day like a junior teammate.

How it helps you

  • Keeps working between check-ins instead of waiting for a prompt.
  • Uses apps that have no API the way a person would, by logging in and clicking through.
  • Reads X in real time, so it catches buyers asking for what you sell as it happens.

Use cases

  • Social listening: watches X for people asking for your service and drafts replies for you to approve
  • Voice: Grok's voice agent answers missed calls, qualifies and books, billed per minute
  • Legacy tools: copies leads from a portal with no API into your CRM
  • Admin: chases invoices, files receipts and updates trackers
  • Prospecting: builds lead lists from X profiles and the web

If you choose it, here is the build

  1. Week 1Accounts and guardrails

    Your xAI plan, separate low-access logins for the bot, and a written list of actions that always need approval.

  2. Week 2Configure the work

    Tasks written in plain English, plus the voice agent if you need calls answered.

  3. Week 3Supervised pilot

    A human reviews every action for one to two weeks.

  4. Week 4Widen and hand over

    More tasks switched on, runbook and training.

The demo you will see

Call the demo number and talk to the voice agent, or watch the bot work one task across three of your apps.

Running cost

Seat-based plans; voice calls are billed per minute (about $0.08 a minute as of November 2026).

What you own

Bots, logins and call logs sit in your xAI account.

Watch-outs
  • Launched in August 2026 and still changing fast, so we start with low-risk tasks.
  • Bots on one account share a computer and logins, so each gets its own low-permission accounts.
  • Outside systems can't easily trigger it yet, so we pair it with Make, n8n or Claude Code when a job must start from an event.
03

The researcher

Perplexity Computer

Perplexity's agent for research-heavy work. It splits a big question across several AI models, cross-checks the answers and hands you a sourced result.

How it helps you

  • Turns hours of research into minutes, with sources you can check.
  • Shared Projects keep your instructions and files, so the team reuses the same research templates.
  • Works where your team already is: web, the Mac app, Slack and Teams.

Use cases

  • Market reports: area comps, price trends and buyer profiles for a listing presentation
  • Sales prep: a one-page dossier on every account before the call
  • Competitors: pricing and positioning scans for agencies and SaaS
  • Content: research briefs that feed your SEO and content agents
  • Investors: partner, investor and vendor due diligence

If you choose it, here is the build

  1. Week 1Seat and Project

    Your Perplexity Max or Enterprise seat, a Project with your instructions and files.

  2. Week 2Research templates

    Reusable skills for your top three recurring questions.

  3. Week 3Connect the output

    Reports delivered to Slack, Docs or your CRM through Composio, on a schedule.

  4. Week 4Handover

    Training on asking good questions and checking sources.

The demo you will see

Name a market, a neighbourhood or a target account. You get a sourced report in a few minutes, on the call.

Running cost

Perplexity Max is about $200 a month and includes 10,000 Computer credits.

What you own

Projects, templates and reports stay in your Perplexity workspace.

Watch-outs
  • It runs on credits, and long multi-agent tasks use more, so we size the plan to your volume.
  • Built for research and reports, not for running your CRM day to day.
04

The rails

Make & n8n

Visual workflow engines. Same steps, every time, with an AI step dropped in only where judgment is needed.

How it helps you

  • Speed: a lead gets a reply in seconds, at 3am or on a Sunday.
  • Reliability: fixed steps, with an alert the moment anything fails.
  • Visibility: your team sees every step on one canvas and can edit it.

Use cases

  • Speed-to-lead: form or portal lead, instant WhatsApp or email reply, qualify, book, log in CRM
  • Reminders: booking confirmations and no-show follow-up
  • Routing: send each lead to the right person or agent
  • Sync: keep CRM, sheets and invoicing in step
  • Triggers: start a Claude Code or Perplexity job when something happens

If you choose it, here is the build

  1. Week 1Map it

    The whole workflow on one page, edge cases included.

  2. Week 2Build it

    Scenario or workflow built and connected to your apps.

  3. Week 3Break it

    Tested with real data, error alerts added.

  4. Week 4Handover

    Recorded walkthrough so your team can edit it.

The demo you will see

We submit a test enquiry. You watch the reply, the booking and the CRM record appear in under a minute.

Running cost

Low monthly plans. Make bills per operation; n8n counts per run and can be self-hosted.

What you own

Scenarios and workflows sit in your Make or n8n account, or on your own server.

Watch-outs
  • We pick Make when your team will edit it, and n8n for high volume, self-hosting or heavier AI steps.
  • Not built for open-ended judgment. That is where an agent step comes in.

By Industry

Which agent we use for your industry, and the demo we show

The outcome comes first. The agent is just how we get there.

Real estate agencies

Make or n8nVoice agentPerplexity

Outcome: Every portal, website and WhatsApp lead gets a reply in under a minute, and no call goes to voicemail.

What we build: Portal or web lead, instant WhatsApp reply, three qualifying questions (budget, area, timeline), viewing booked in the agent's calendar, CRM record with the full chat. Missed calls go to a voice agent (Grok voice, Retell or Vapi). Perplexity builds area reports for listing presentations.

The demo: We submit a test enquiry live on the call. You watch the WhatsApp reply, the qualifying chat, the booked viewing and the CRM record appear in about a minute.

Marketing agencies

Claude CodePerplexity

Outcome: More content shipped and every client report done on time, without adding headcount, plus a steady new-business pipeline.

What we build: A content agent that writes in each client's brand voice from their brief, a Monday client report pulled from ads and analytics, and an outbound agent for your own new business. Perplexity writes the research briefs behind the content.

The demo: Give us one client brief. Claude Code drafts a week of posts in that brand's voice and a one-page performance report, live.

Coaches & consultants

MakeClaude Code

Outcome: No lead goes cold while you are in a session, and a fresh list of people to message every morning.

What we build: Make replies to every enquiry and DM with a booking link within a minute and chases no-shows. Claude Code runs the SDR lead engine: new followers of pages your buyers follow, filtered and researched, with three personal openers each.

The demo: Fill in the demo form or send a DM and get a reply and booking link within a minute. Then see this morning's lead sheet with openers written for each person.

B2B SaaS

Claude CodePerplexityGrok Bot

Outcome: A daily outbound pipeline without hiring an SDR team, and content that AI search tools cite.

What we build: Claude Code with Composio runs signal-based outbound: find accounts, research them, write personal emails, log everything in your CRM. Perplexity builds account dossiers for sales calls. Grok Bot watches X for buyers asking for tools like yours.

The demo: Name one target account. You get a research dossier, three personalised emails and a CRM entry in a few minutes.

B2B service businesses

Voice agentMake or n8nClaude Code

Outcome: Every call answered, every lead followed up the same hour, every proposal out the same day.

What we build: A voice agent answers and qualifies calls and books meetings, Make or n8n handles reminders and CRM updates, and Claude Code turns the call recording into a branded proposal deck.

The demo: Call the demo number. The agent answers, qualifies you and books a slot. Then we upload a call recording and you get a proposal deck back.

Questions founders ask

Which agent will you use for my business?

The one that fits the job, and often two. We map your week on the strategy call, then pick: Make or n8n for fixed pipelines, Claude Code for judgment and writing, Perplexity for research, Grok Bot for always-on work and X. You get the reasoning in writing before we build.

Can we switch agents later?

Yes. Your app connections run through Composio, so they carry over when you move a job from one agent to another. The playbook and skills are plain text you own.

Do we need technical staff?

No. Day to day, your team approves drafts in Slack or reads a report. We handle the build, and the handover covers what to edit and what to leave alone.

Who owns everything?

You do. Every agent, workflow and connection is built in your own accounts under your logins. If we part ways, it keeps running.

Do you use Clay?

No. For complex, multi-signal outbound builds on Clay we refer you to RevPartners, Clay's official Elite Studio Partner. Everything we build runs on the agents on this page.

Supporting Tools

The tools we plug into these agents

Data, email, LinkedIn, voice and booking tools the agents use to do the work. Open a category to see what we use and why.

Full disclosure: Six Figure Consulting is an affiliate partner for most of these tools. If you sign up through one of these links we may earn a commission, at no extra cost to you. We only list tools we use in our own systems.
GTM Orchestration5 tools

The workflow and automation layer that ties every other tool in this stack together.

  • MakeThe no-code automation backbone connecting most of our client tech stacks.
    Make →
  • n8nThe open-source workflow engine we run alongside Make to orchestrate outbound, without the expensive per-seat pricing of tools like Clay.
    n8n →
  • CampaignFlow ProOur own campaign orchestration platform, built in-house to run and manage outbound at scale.
    CampaignFlow Pro →
  • DeeplineGTM data infrastructure built for AI agents: waterfall enrichment, lead and account discovery, and signal detection, run through a Claude Code skill instead of a static contact database.
    Deepline →
  • Claude CodeThe AI coding agent we run our GTM workflows through directly, including Deepline's enrichment and outbound plays, not a chatbot bolted onto the side.
    Claude Code →
Data Enrichment & Web Scraping9 tools

Where our lists, enrichment waterfalls, and scraping pipelines start: the data that makes personalization possible.

  • FindymailFinds and verifies decision-maker emails so your lists convert instead of bouncing.
    Findymail →
  • TheirStackTech stack and job posting data for precise, intent-based account targeting.
    TheirStack →
  • ExpleeAI-powered lead search that finds companies and decision-makers matching your ICP from a plain-English description.
    Explee →
  • LushaContact and company data to fill the gaps other providers miss.
    Lusha →
  • Bright DataEnterprise-grade proxies and scraping infrastructure powering our data pipelines.
    Bright Data →
  • ScraperAPIHandles the proxy and CAPTCHA headaches so our scrapers just run.
    ScraperAPI →
  • ApifyPre-built and custom scrapers for pulling structured data from almost any site.
    Apify →
  • Phantom BusterCloud automations for scraping and outreach across LinkedIn, Instagram, and more.
    Phantom Buster →
  • BoomerangAutomated LinkedIn engagement data that keeps your profile visible before you ever reach out.
    Boomerang →
Cold Email & Deliverability3 tools

The infrastructure and verification layer that keeps sending domains warm and lists clean.

  • ZapmailDone-for-you inbox infrastructure so your sending domains stay warm and out of spam.
    Zapmail →
  • BouncifyEmail verification that keeps bounce rates low and sender reputation clean.
    Bouncify →
  • ZerobounceA second layer of email validation we run before any list goes live.
    Zerobounce →
AI Cold Email Tools2 tools

The AI-assisted sending platforms we actually run outbound campaigns on.

  • LemlistMultichannel outreach that combines email, LinkedIn, and calling in one sequence builder.
    Lemlist →
  • InstantlyCold email sending infrastructure with built-in deliverability: our go-to for scaling outbound safely.
    Instantly →
LinkedIn Automation & Outreach3 tools

The two LinkedIn platforms we trust enough to run our own outreach on, plus the AI copilot that speeds up the research behind it.

  • GetSalesLinkedIn automation for connection requests, messaging, and CRM sync in one place.
    GetSales →
  • HeyReachMulti-account LinkedIn outreach built for agencies running outbound at scale.
    HeyReach →
  • GojiberryAI browser copilot that automates repetitive LinkedIn research and prospecting tasks.
    Gojiberry →
AI Sales Agents & Personalization4 tools

AI that does the outreach work directly: voice calls, video, inbox agents, and personalization at a scale no human team can match.

  • ReeonAI voice agents that call, qualify, and book meetings on autopilot.
    Reeon →
  • DeviAn AI sales agent that handles outbound conversations so your team doesn't have to.
    Devi →
  • RepliqHyper-personalized video and landing pages generated at scale for cold outreach.
    Repliq →
  • InstantlyInstantly's AI sending agent handles inbox management and reply triage on autopilot.
    Instantly →
Voice AI & Calling3 tools

The platforms we build client voice agents on, for customer support and outbound to a client's own lead list, never cold calling.

  • VapiA developer platform for building production voice AI agents, one of the two stacks we build client agents on.
    Vapi →
  • Retell AIA voice agent platform built for real-time phone conversations, the other stack we build client agents on.
    Retell AI →
  • ElevenLabsNatural-sounding AI voice generation that powers how our client voice agents actually sound.
    ElevenLabs →
AI Productivity & Research4 tools

The tools we personally run every day to move faster on calls, research, writing, and busywork.

  • FathomFree AI notetaker that records, transcribes, and summarizes every client call.
    Fathom →
  • PerplexityOur default AI search engine for fast, cited research on anything.
    Perplexity →
  • Manus AIAn autonomous AI agent that can plan and execute multi-step tasks end to end.
    Manus AI →
  • ClaudeThe AI assistant we use daily to write, build, and run this entire tech stack.
    Claude →
Automation & CRM3 tools

Where the workflows live and where the pipeline gets tracked once leads start replying.

  • ALM (Automatic Leads Machine)Done-for-you lead generation automation for agencies and consultants.
    ALM (Automatic Leads Machine) →
  • PipedriveA simple, visual CRM pipeline for teams that don't need Salesforce-level complexity.
    Pipedrive →
  • BreakcoldA social-selling CRM built around LinkedIn and Twitter engagement, not cold calls.
    Breakcold →
Content, Community & Growth4 tools

For building an audience and a network around your business, not just a pipeline.

  • SkoolThe community platform we'd choose to run a paid course or membership on.
    Skool →
  • BeehiivThe newsletter platform behind our own AI & GTM Weekly.
    Beehiiv →
  • SupergrowAI-assisted LinkedIn content writing to grow a personal brand consistently.
    Supergrow →
  • Our LinkedIn GroupJoin our free LinkedIn community of founders and GTM operators sharing what's working.
    Our LinkedIn Group →
Financing Partners1 tools

For clients who need capital to say yes to an opportunity today.

  • Closer CapitalWorking capital and funding options for growing businesses and agencies.
    Closer Capital →
Privacy & Infrastructure4 tools

The unglamorous tools that keep everything else running securely, and the same production stack we build client websites on.

  • ProtonEncrypted email and privacy tools we use to keep sensitive business data secure.
    Proton →
  • VercelProduction hosting for this site and every website we build for clients: fast, reliable, zero server management.
    Vercel →
  • GitHubVersion control for every site and system we build, so you own the code, not just the live pages.
    GitHub →
  • CloudflareDomain, DNS, and CDN for fast, secure delivery of every site we host.
    Cloudflare →
AI Images & Video Creation2 tools

The tools we use to generate and animate visuals for content and campaigns.

  • OpenArtAI image generation for concepts, product visuals, and campaign creative.
    OpenArt →
  • HeyGenAI avatar and video generation for scaling content without a camera crew.
    HeyGen →
Project Management3 tools

Where the work itself gets organized, documented, and tracked day to day.

  • NotionOur docs, wikis, and project trackers, all in one flexible workspace.
    Notion →
  • Google WorkspaceEmail, docs, and file storage that keep our team and clients on the same page.
    Google Workspace →
  • ClickUpTask and project management for running client work and internal ops in one place.
    ClickUp →
Scheduling & Booking1 tools

How we and our clients actually get calls on the calendar.

  • Cal.comOur preferred booking tool. It connects to Claude through MCP, so our AI teammate can check availability and book calls directly, and it collects payment at the time of booking, not after.
    Cal.com →
Find Talent to Run Your Business2 tools

Free communities where founders find appointment setters, closers, and GTM operators.

  • Facebook GroupJoin our free Facebook community of 7,000+ members to find appointment setters and outbound talent.
    Facebook Group →
  • LinkedIn GroupJoin our free LinkedIn community of 1,000+ founders and GTM operators sharing what's working.
    LinkedIn Group →

Not sure which agent fits?

Tell us about your week. We’ll tell you which agent to build first, what it will cost to run, and what you’ll see in the demo.