Difference between output and result
Over the past two years, I’ve had hundreds of conversations with business leaders about AI. The pattern is always the same: initial excitement about what AI can do, followed by disappointment at what it actually delivers.
“It writes great emails,” he tells me. “But are they changing?” Not really.
“It can research possibilities in seconds.” very good But is it looking right? Not usually.
The gap between AI’s impressive output and the business results we actually need has never been clearer. At HubSpot, we have an opinion on why this gap exists. That’s because AI today operates without one thing that humans have always brought to their work: context.
The missing ingredient: context
Think about the best members of your team. They know which possibilities are worth prioritizing because they have seen similar patterns before. They write emails that convert because they understand your brand’s voice and your audience. They solve customer problems quickly because they remember what worked in the past.
This context: getting the right information at the right time, combined with the decision to know what to do with it.
Without context, AI gives generic answers based on the internet’s one-size-fits-all wisdom. It doesn’t know what the actual results are for your customers, your business, or companies like you. This is why AI agents without context can’t find the results you’re looking for. They lack the understanding needed to make their actions effective.
Problem: Context lives everywhere and nowhere
Right now, most of this perspective is buried in people’s heads, scattered across systems, and hidden from how your team actually works.
Why did you extend this contract? Why did this campaign work? Most systems capture what happened, but the reasoning behind the decisions (exceptions, precedents, judgment calls) is spread out over email, slow and lives in people’s heads.
This inner knowledge grows over time. And when someone leaves, it disappears.
Most AI tools cannot find or use this context. They ask you to upload your brand guidelines, configure your workflows, and teach them how your business works. Then repeat for the next device. And then
It’s backwards.
What we’re building: Agentic customer platform
At HubSpot, we believe there is a better way. What we call an agent customer platform: software that captures all of your customer data and business context in one place, then makes it available to both your team and AI agents so they can work together to market, sell and service customers.
For over 25 years, CRM software has helped track the work of marketing, sales and support teams. The software contained the data, and humans provided the context, judgment and action. Now, AI agents can do more, but only if they have access to the same perspective as humans.
This requires a fundamentally new type of platform, built on three tightly coupled layers.
1. Contextual Layer: Where the customer perceives life
At its core is our SmartCRM, a single source of truth that brings together:
Complete customer data: Not only structured records like companies, contacts, and tickets, but also unstructured data in emails, call transcripts, and chat conversations. The whole picture, all in one place.
Business Context: Your unique brand, products and strategy. Not just what happened, but why the decisions were made. Understanding rationality, precedent, exceptions, and judgment calls.
Team Context: How your people work, communicate and collaborate. The patterns that make your team effective are systematically captured so that knowledge is not lost when someone leaves.
Industry Intelligence: Insights from over 250,000 companies and their customers. What works for a business like yours: Which campaign tactics drive results, which deal indicators predict closes, which keep customers happy.
Domain Knowledge: Over 20 years and deep marketing expertise capturing millions of campaigns, deals and support interactions. It guides smart actions such as when a deal needs attention today, when a prospect is ready to buy, or how to route a complex issue.
2. Action Layer: Where the work is done
Context alone is not enough. You need applications that can actually implement this to drive value:
Hubs: Our marketing, sales, and service centers are smart from day one because they are contextual. This makes it much easier to grow awareness, build pipelines, and keep customers happy.
Air agents: AI companions who do real work: researching accounts, enriching data, answering support questions, qualifying leads. They complete tasks that free up your team for higher-value work.
Air Assistant: An AI-driven market expert for each employee, providing insights and recommendations relevant to their role and context. It can also write and update CRM for you.
What makes them different? They are informed by the full context of the customer, so they create better results for your business and better experiences for your customers. And because it’s all on one platform, you avoid the spam and disconnected experiences that come with fragmented point products.
3. Coordination layer: where humans and agents collaborate
For humans and agents to work as true partners, you need synergy that goes beyond traditional workflow automation.
Agent Management: You decide which tasks agents handle autonomously and which remain with humans. You can @manage agents in deals, assign them tasks, set permissions (just like an employee).
Linked System: Your business doesn’t run in one system, and neither do your agents. HubSpot connects to your broader tech stack so agents can work across platforms, which become the connective tissue between systems.
Unified Governance: A platform means a set of permissions, an audit trail, a security model. Everything is governed by constant rule.
Why does it matter?
You don’t care about AI. You care about results. And results need context.
AI models are becoming commoditized. What AI can’t replicate is your combination of customer data, business knowledge, and proven practices from thousands of similar companies. Ai knows a lot about the world. You need a platform that you know about yours The world
Most CRMs were not designed for this. They are systems of record, not systems of context. AI Point products do not have access to your complete customer data. And integrating together creates the very fragment that the shared context is meant to resolve.
With HubSpot, your context lives in one place and powers everything. Where humans and AI collaborate as teammates. Where you can put your customers at the center of market, sales and service.
This is our vision: to make AI actually work for growing companies.
