5/28/25: The GTM Shift in the Age of AI Agents, Interfaces, and Protocols
This newsletter is sponsored by AI Business Network and the GTM AI Academy
AI Business Network is a community for business leaders in finance, revenue, HR, accounting, legal or other departments to have insight on how Fortune 500 companies and startups are using AI effectively.
GTM AI Academy is for GTM professionals who want to learn how to apply AI in their day to day jobs and make an impact on business.
This week we have the GTM AI Podcast with Scott Martinis of B2B Catalyst
With some articles below we will deep dive into and find patterns:
Source Articles Summary with Links
SEO is Dead, According to Google – Inc. argues that Google's shift to AI-first search results eliminates traditional SEO, urging businesses to create meaningful, unreplicable content and lean into branded, human-centered engagement.
OpenAI CFO on Acquisition of Jony Ive’s Startup – OpenAI reveals plans to integrate bespoke AI hardware to spark a “new era” of computing and subscriptions, expanding user interaction beyond the phone.
Google's AI Mode Shopping and Virtual Try-On – Google is transforming ecommerce through AI-powered shopping experiences, agentic checkout, and personalized virtual try-ons.
OpenAI ChatGPT to Integrate MCP – ChatGPT will support the Model Context Protocol (MCP), allowing enterprises to plug in internal tools and data into ChatGPT workflows.
Microsoft Multi-Agent Collaboration Announcement – Microsoft launches enterprise-grade multi-agent systems in Copilot Studio, enabling collaborative task handling and agent interoperability across platforms.
Lets get into the podcast!
Breaking Down GTM Engineering with Scott Martinez: A Game-Changing Conversation
You can go to Youtube, Apple, Spotify as well as a whole other host of locations to hear the podcast or see the video interview.
I just had one of those conversations that makes you want to completely rebuild your entire go-to-market motion. Scott Martinis of B2B Catalyst dropped some absolute truth bombs that I'm still processing.
Let me be straight with you - I've been in sales and enablement for years, and Scott's approach to GTM engineering is unlike anything I've seen. This isn't your typical "send more emails" or "hire more SDRs" playbook. This is surgical precision applied to revenue generation.
The Big "Aha" Moment
Scott shared a story that stopped me in my tracks. He generated 700 MQLs across three companies - 180 for one, 90 for another, and 399 for the third. Guess how much converted to revenue? Zero. Zilch. Nada.
Why? Because generating leads isn't the same as generating revenue. And that's where most of us get it wrong.
The Account Qualification Revolution
Here's what blew my mind: While most RevOps teams are doing territory planning based on industry and company size, Scott's data shows that proper account qualification criteria can result in 2-5x higher close rates.
Think about that. If you're targeting accounts outside your true ICP, you're operating at 50-80% reduced effectiveness. You could make 100 calls into qualified accounts and get 5x better results than the same effort into unqualified accounts.
The Smart GTM Playbook (What Scott Actually Does)
Step 1: Find Your True ICP Through Forensic Analysis
Interview your top 3 sales reps with a "Perfect Opportunity Worksheet"
Ask them: "When you're researching the best prospect ever, what do you expect to see?"
Look for specific signals:
Do they have 24/7 support? Chat support? Email support?
What team structures exist? (RevOps + Sales Enablement = 2x more likely to close)
What's on their website? Demo buttons? Support centers?
Department headcounts and trends
Step 2: Fix the Bottleneck, Not Everything
Scott's approach is brilliant here. Instead of trying to automate everything at once, he asks: "What's the one constraint that, if fixed, would unblock everything else?"
Real example: An SDR team spending 2 hours per day on account qualification. Instead of replacing them with AI, Scott's team:
Identified 13 discrete website signals
Built a scoring rubric
Automated the qualification process
Ran 80% of their CRM through it
Found all the whitespace in their market
Result? SDRs got 2 hours back per day, marketing got proper targeting, and AEs could finally hit self-sourcing targets.
Step 3: The Funnel Math That Actually Works
Here's the exact math Scott uses (and you should too):
To hit $10M ARR:
Need: 180 new customers at $50K each
At 25% close rate = 720 opportunities needed
At 20% meeting-to-opp rate = 3,600 meetings needed
At 20% conversation-to-meeting rate = 18,000 conversations needed
At 20% contact-to-conversation rate = 90,000 dials/emails needed
With 5 contacts per account = 18,000 accounts needed
But here's the kicker - every 10% of unqualified accounts in this mix torpedoes your downstream metrics.
The AI Reality Check
Scott's take on AI is refreshingly practical: "AI on its own is useless. You have to target it, constrain it, focus it, and give it examples to mimic and scale."
His process:
Understand the manual process that works
Document exactly how your best people do it
Use AI to scale that proven process
Never try to AI your way around a broken process
The Tools That Matter (And Why)
Scott doesn't worship tools, but he's specific about what works:
Phone data: You need 20%+ connect rates. If you're at 3%, your data sucks
Email: Industry average is dying. Apollo worked a year ago, doesn't now
Clay: Great for enrichment, but it's <50% of the actual work
Dialer stack: Get your team having 3-5 conversations per hour
The Metrics That Actually Matter
Forget activity metrics. Here's what to track:
Qualified account identification rate
Contact-to-conversation rate (aim for 20% with good data)
Conversation-to-meeting rate (10% minimum, fix messaging if lower)
Meeting-to-opportunity rate
Close rate by account qualification score
My Biggest Takeaway
After years in sales and enablement, I thought I understood pipeline generation. But Scott's approach made me realize we've been playing checkers while he's playing 3D chess.
The difference? We try to do more. Scott figures out what actually works, then scales it intelligently.
Your Action Items:
This week: Run the Perfect Opportunity Worksheet with your top 3 reps
Next 30 days: Identify your #1 GTM bottleneck (hint: it's probably account qualification)
Next quarter: Build ONE automated process that unlocks your biggest constraint
Ongoing: Track the right metrics religiously, but use them to guide investigation, not make decisions
Look, I've sat through hundreds of GTM conversations. This one's different. Scott's not selling snake oil or the latest shiny tool. He's showing us how to think about go-to-market as an engineer would - systematically, measurably, and intelligently.
If you're not following Scott Martinez on LinkedIn yet, you're missing out on some of the most tactical, practical GTM content out there. This conversation fundamentally changed how I think about pipeline generation.
Time to stop throwing bodies at the problem and start engineering solutions that actually scale.
From Interface to Infrastructure: The Emerging GTM Landscape in an Agent-First Economy
Introduction:
The convergence of developments from OpenAI, Google, Microsoft, and third-party agents is no longer speculative. It is systemic. We are now observing a market-wide structural shift: search as an interface is collapsing; SEO is disintegrating under AI summarization; agentic systems are bypassing traditional customer acquisition flows; and product experiences are moving from apps to protocols.
GTM leaders must now adapt their strategies to a new terrain governed by context-rich agents, persistent AI interfaces, multimodal interaction, and zero-click commerce. The "funnel" is now a mesh of agent workflows. The user is not your buyer — their agent is. And being discoverable is no longer enough. You must be "actionable" in an agent-native economy.
1. The Collapse of SEO: From Search Visibility to Feed-Level Integration
The death of SEO is not hyperbole. According to Google's own announcements at I/O and recent market data:
AI summaries now occupy the top of nearly all search queries, replacing traditional blue-link hierarchies.
Inc. reports that traditional SEO-based discoverability has declined by 400% for some businesses, while anti-SEO strategies — focusing on branded, irreducible content — have increased traffic by the same magnitude.
Google’s AI Mode and Shopping Graph now reshape search into action-led workflows:
AI Mode queries now fan out simultaneous sub-queries, scan Shopping Graph data, and return product recommendations without a user ever viewing traditional listings.
The Shopping Graph refreshes 2 billion+ listings per hour across 50+ billion product SKUs.
What this means for GTM teams:
Your top-ranking content is now either invisible or summarized by a model that may misrepresent your brand.
Branded, unreplicable content (e.g. personalities, newsletters, live commentary) is the only SEO-aligned content AI won’t easily replace or compress.
Organic search strategy must now be replaced by agent feed strategy. Think: "What structured product information, reviews, pricing, and images will an agent need to recommend us?"
Actionable GTM Adjustments:
Migrate SEO budgets to product feed optimization and agent content kits — structured JSON-based assets that agents can parse.
Build visibility into how your brand is represented in AI summaries by different platforms (Google, ChatGPT, Perplexity).
Launch branded communities and newsletters as primary discovery surfaces — SEO is no longer reliable top-of-funnel real estate.
2. Agents as the New Operating System for Consumer and B2B Journeys
We have now passed the interface phase of AI. Agents are no longer just wrappers or copilots — they are replacing apps and browser-based flows as the default mode of task completion. Across Microsoft, OpenAI, and Google, AI agents are now capable of:
End-to-end checkout (Google Shopping’s “buy for me” flow),
System-wide orchestration (Microsoft’s new multi-agent Copilot Studio architecture),
Contextual task routing via Model Context Protocol (MCP) (OpenAI, HubSpot, Claude).
In Microsoft’s rollout of Copilot multi-agent architecture:
A single process — such as proposal generation — is now handled by a network of collaborating agents: one pulling sales data, one drafting the proposal, another triggering a calendar invite.
These agents operate across systems — CRM, Outlook, SharePoint — without requiring user clicks or manual inputs.
Microsoft also released agent discovery routing, where users simply describe a task and are routed to the appropriate agent using vector-based matching.
OpenAI’s adoption of MCP pushes this even further:
MCP enables plug-and-play integrations with internal business tools, allowing ChatGPT to perform CRM tasks, database queries, or pipeline updates without bespoke APIs.
HubSpot’s MCP server beta already enables agents to create deals, summarize pipelines, and log CRM entries with simple natural language prompts.
What this means for GTM teams:
Users will no longer visit your app to complete tasks. Their agent will.
Most buyer interactions — particularly in B2B — will be initiated, qualified, and routed by agents operating on behalf of decision-makers.
The buyer is no longer your entry point. Their preferred agent is — and your product must be callable via that agent.
Actionable GTM Adjustments:
Implement MCP support or structured APIs for integration into agent environments like ChatGPT, Claude, and Copilot.
Treat each GTM task (demo booking, case study routing, renewal updates) as agent-compatible workflows, not button clicks.
Build GTM systems that support multi-agent orchestration: one agent for qualification, another for product recommendations, another for pricing negotiation.
3. Agent Swarms and Multi-Agent Systems: The New Funnel Infrastructure
In traditional GTM models, the funnel has discrete, human-owned stages: MQL, SQL, Demo, Close, Retention. In the emerging GTM model, each stage becomes a micro-agent or swarm capable of autonomous or semi-autonomous action.
Microsoft now supports agent swarms that collaboratively execute business processes:
For example, scheduling follow-ups based on prior performance data, drafting proposal text using fine-tuned models, and managing data handoffs through RAG + vector architecture.
This architecture now supports custom model selection and Python-powered code interpreters — allowing agents to reason, self-correct, and present validated financial analysis or GTM forecasts.
Google and OpenAI are also converging on this architecture:
Google's AI Shopping Mode agents use query fan-outs and parallel reasoning paths to identify optimal product recommendations.
OpenAI's leaked MCP integration shows plans for agents to call custom tools and connectors, enabling fine-grained enterprise workflows.
This implies that agents will no longer be isolated assistants. They will form collaborative, cross-functional units that replicate internal teams.
What this means for GTM teams:
A new operating model is emerging: instead of SDRs booking demos, agents qualify leads, route to a demo agent, and automatically log results in CRM.
Instead of CSMs managing renewals, agents pull usage data, compare benchmarks, and generate retention playbooks in real time.
The “stack” is becoming the org chart — with agents representing key workflows, not just software modules.
Actionable GTM Adjustments:
Reframe your GTM architecture as a set of agents, not functions.
SDR Agent: Lead qualification, email routing.
Demo Agent: Custom walkthroughs and integration validation.
CS Agent: Renewal workflows and usage-based playbooks.
Build an “Agent QA” function inside Enablement or RevOps. Agents need onboarding, performance reviews, escalation protocols — just like employees.
Define agent lifecycle metrics: task accuracy, escalation rate, workflow completion time, and customer satisfaction with agent output.
From Interface to Infrastructure: The Emerging GTM Landscape in an Agent-First Economy
4. AI Hardware Will Reshape How Products Are Accessed, Not Just Where They're Used
OpenAI’s $6.4 billion acquisition of Jony Ive’s startup signals a deeper shift: GTM is about to move from screen to substrate. Devices aren’t just passive delivery surfaces anymore. They're becoming intelligent agents in themselves — voice-first, screen-optional, multimodal, always-on.
Key insights from the OpenAI hardware strategy:
CFO Sarah Friar confirmed that custom AI-first hardware will drive higher ChatGPT subscription attach rates by embedding AI agents directly into user behavior.
Devices being considered move beyond touchscreens and are focused on natural human interfaces — voice, gesture, context.
The goal is to build persistent, integrated agent presences that operate continuously in a user's life, not just when an app is opened.
These dynamics parallel previous platform shifts:
Mobile changed where interactions happened.
AI hardware will change how and when those interactions begin — often with no explicit trigger from the user.
What this means for GTM teams:
Expect to compete for user mindshare via hardware-level agent presence, not just app installations or web traffic.
The operating system of the next computing era won’t be macOS or Android — it will be the agent layer embedded in hardware.
If OpenAI succeeds in launching hardware with ambient ChatGPT, the first touchpoint with users will be agent-driven by default.
Actionable GTM Adjustments:
Begin developing product flows that can be triggered by voice or zero-click contextual cues (e.g., "what’s the next step for Acme Co?" instead of opening your dashboard).
Track agent distribution platforms (Meta Ray-Ban glasses, OpenAI hardware, voice-first interfaces) and plan for new GTM channels through these interfaces.
Integrate subscription strategies with agent usage data — the next conversion metric isn’t just page views, it’s agent prompt frequency.
5. Shopping and Commerce Are Becoming Agentic, Not Navigational
Google’s AI Mode for Shopping is a case study in how agents are reshaping consumer commerce:
Users can now initiate intent ("find a travel bag"), and the AI fans out sub-queries, filters through 50B+ product listings, and presents visual, contextual, personalized options — all before the user clicks anything.
Checkout is agent-led. Users tap “track price,” then hit “buy for me.” Google agents handle cart insertion, checkout, and secure payment autonomously.
Their virtual try-on feature lets users upload a photo and try on billions of clothing SKUs with real-time image generation.
These changes transform the purchase path:
It no longer involves navigation, filters, or comparison charts.
It becomes a single agent-driven decision loop from inspiration to action.
What this means for GTM teams:
Your PDP (product detail page) is no longer a decision point — it’s a dataset.
Agents don’t need marketing copy; they need structured, high-quality product metadata, customer reviews, and supply chain integration.
Discovery is now behaviorally routed — not keyword-triggered.
Actionable GTM Adjustments:
Structure your entire product catalog for agent interpretation (title, price, availability, specs, images, tags).
Build agent-readiness KPIs for your ecommerce stack: Are your products usable in Google Shopping Graph, ChatGPT plugins, or AI-powered discovery tools?
Shift marketing from CRO (conversion rate optimization) to AIO — Agent Interaction Optimization: fewer steps, richer data, seamless triggers.
6. Discovery Is Now Protocol-Driven, Not Platform-Driven
The integration of Model Context Protocol (MCP) into ChatGPT, Claude, HubSpot, and other platforms redefines discoverability. Agents no longer discover products by search. They integrate products through APIs, context endpoints, and protocol connectors.
Key signals:
ChatGPT is testing “Add Connector” functionality to directly plug in enterprise apps, custom tools, and databases.
Once connected, tools can be invoked contextually by the agent — without ever being listed on a marketplace or shown in a search result.
This inverts the old platform paradigm:
You no longer fight for ranking on a platform’s surface.
You fight for relevance inside an agent’s contextual workflow — and that’s governed by protocols, not interfaces.
What this means for GTM teams:
Integrations matter more than listings. If you’re not callable via MCP or similar protocols, agents won’t use you.
Marketplaces will evolve from static app stores to dynamic toolkits inside agent prompts.
Contextual exposure is now a GTM capability — not just a product feature.
Actionable GTM Adjustments:
Assign an “Agent Integration Owner” in your GTM org — someone responsible for MCP endpoints, structured tool docs, and AI readiness.
Treat protocols like you treat SEO — build for discoverability inside agent tool call architectures.
Develop “tool cards” — natural language descriptors of what your tool can do, how, and under what conditions, so agents can self-discover and invoke your product.
7. GTM Orgs Are Becoming Agentic Systems Themselves
Microsoft’s Copilot Studio and OpenAI’s evolving orchestration frameworks show that companies are no longer just building tools. They are building agent-native business units.
Across enterprise rollouts, we now see:
GTM workflows are being mapped to agents: pre-sales, CS, renewals, pricing, enablement.
Microsoft’s multi-agent orchestration allows for task delegation across agents, replacing internal cross-functional meetings with programmatic execution.
Admins can now discover, route, and manage agents like team members — including permissions, approvals, and audit trails.
This is not an incremental change in tooling. It’s a redesign of the human-machine org chart.
What this means for GTM teams:
Your org design must now include agent flows as first-class citizens.
Enablement needs to include agent design, prompt QA, and lifecycle management.
Reporting lines will shift: some outputs will be owned by agents, overseen by human managers.
Actionable GTM Adjustments:
Create a centralized “AgentOps” function in RevOps or BizOps to own:
Agent prompt libraries
QA protocols
Escalation paths
Workflow monitoring
Redesign GTM playbooks into modular agent workflows, with humans supervising outcomes, not executing steps.
Track agent contributions to key GTM KPIs (pipeline created, demos booked, contracts renewed).
Closing Outlook
The old playbook was interface-first, with humans at every decision node. The new playbook is protocol-first, with agents executing flows autonomously, invisibly, and instantly.
GTM leaders now need to restructure around the following realities:
Your “users” are agents.
Your visibility is mediated by protocols.
Your conversion flows are orchestrated across swarms.
Your differentiation depends on how callable, usable, and trustworthy your product is in an agent-first economy.
Every buyer journey is becoming a mesh of agent transactions.
The GTM organization of the future is not a team using AI. It is a system of humans and agents co-operating to deliver outcomes. Those who adapt first will not just accelerate revenue. They will redefine how software is marketed, discovered, and consumed.
GTM AI Tool of the Week: Tidio
The AI-Powered Customer Experience Platform You Might Be Overlooking
Tidio is quietly becoming a sleeper hit among small to mid-sized businesses—and increasingly, even larger GTM teams—looking to modernize how they engage, convert, and retain customers through AI. If your GTM motion includes digital engagement, service automation, or customer support at any touchpoint, Tidio may deserve your attention.
What is Tidio?
Tidio is an all-in-one customer experience (CX) platform built for businesses that want to combine live chat, helpdesk, chatbot automation, and AI into a unified inbox. Think of it as an AI-enhanced frontline customer ops layer. Its AI assistant “Lyro” helps automate up to 70% of customer service inquiries, while the platform itself integrates with CRMs, ecommerce platforms (Shopify, WooCommerce, BigCommerce), and marketing stacks with ease.
It's designed to allow businesses to manage customer communication at scale without bloated tech stacks or massive support teams.
Core Capabilities
1. Lyro AI Agent
Lyro is Tidio’s conversational AI engine that can answer repetitive customer questions, learn from your knowledge base, and escalate issues to humans when needed. It’s similar to Intercom’s Fin or Zendesk’s Answer Bot, but designed with SMB accessibility in mind.
2. Live Chat + Unified Inbox
Agents can manage conversations from email, Messenger, Instagram, WhatsApp, and live chat—all in one place. This is especially useful for lean GTM teams that need centralization.
3. Helpdesk Ticketing
Basic support ticketing features baked into the inbox, with integrations into external CRMs or support stacks.
4. Lead Generation Bots
Chatbots can be used to qualify leads, push promo offers, capture emails, and even offer product recommendations—all without code.
5. Deep Integrations
Native integrations with HubSpot, Shopify, WordPress, Mailchimp, Google Analytics, and more.
Strengths
Ease of Setup: Deploy AI chatbots and live chat in minutes—no engineering required.
Affordability: Compared to Intercom, Zendesk, or Drift, Tidio is significantly less expensive, with premium AI features bundled in.
Focused on Outcomes: Lyro promises to automate up to 70% of support queries, reducing ticket volume and headcount.
Built-in Growth Feature: Lyro promises to automate up to 70% of support queries, reducing ticket volume and headcount.
Built-in Growth Features: Lead gen CRM syncing, behavior-triggered chat, and chatbot funnels, GTM teams can drive more pipeline without hiring more reps.
ROI Potential
The economics are clear:
A typical customer support rep costs \$40K–\$70K/year.
Tidio’s Lyro plan starts around \$39/month per agent and can automate up to 70% of inbound support traffic.
Use cases like abandoned cart recovery and promo bots have demonstrated conversion boosts up to 30% on Shopify stores using Tidio.
That’s immediate ROI in:
Fewer tickets → Lower support costs
24/7 availability → Faster response times and higher CSAT
Proactive lead capture → More qualified pipeline, especially in B2C GTM teams
Conversion-focused bots → Increased online sales with zero human input
Who It’s Best For
SMB GTM Teams that need to scale support and lead gen without headcount.
Ecommerce Brands using Shopify, WooCommerce, or Wix.
Startups who want AI-driven CX without investing in a \$50K/year Intercom license.
GTM Leaders looking for fast win pilots in customer engagement or support deflection.
Strategic Consideration
Tidio represents the direction many GTM leaders are heading: AI-first, low-ops tools that reduce manual handoffs, increase conversion rates, and keep cost structures lean. While not as advanced as full-stack AI agent orchestration platforms, it’s a powerful entry point for teams who want to use AI to create better customer experiences.
As more buyers expect instant, accurate, and 24/7 responses, tools like Tidio are becoming not just “nice to have” — but expected. And for GTM professionals, that means more efficiency, higher CSAT, and more revenue—without the bloat.