Most outbound sales motion looks the same: buy a list, blast emails, hope for the best. Conversion rates hover around 1–2%, reps burn out, and the pipeline looks thin. The problem isn't outbound, it's undirected outbound. Reaching out to companies with no signal is a volume game at best, and a brand-damaging exercise at worst.
HubSpot has quietly assembled a set of tools that fundamentally change this dynamic. When used together - Target Markets, Buyer Intent, Research Intent, and the AI Prospecting Agent - they create a layered signal engine that lets you prioritize outreach toward the companies most likely to buy, and personalize it at scale. This post walks through exactly how to set that up.
The article explains that effective outbound prospecting isn’t about volume, but about timing and relevance: signals first, then outreach. With HubSpot, you can approach this by defining target markets based on your ICP and technographic filters, then detecting buying intent through Buyer Intent (website visitors) and Research Intent (companies researching relevant topics), and prioritizing these signals in tiers.
Next, you enrich only the most valuable companies with contact data using tools like Apollo.io and use the AI Prospecting Agent to generate personalized outreach at scale. By combining this end-to-end workflow and continuously optimizing it based on response and data quality, outbound shifts from an inefficient “spray-and-pray” model to a data-driven, largely automated signal machine with significantly higher conversions and stronger pipeline generation.
The old model was: define your ICP broadly, export a contact list from Apollo or ZoomInfo, and start reaching out. The new model inverts this. Before you even think about reaching out, you want to know:
HubSpot's prospecting stack maps directly to each of these four steps.
The foundation is HubSpot's Target Markets feature (found under Settings → Markets). This lets you define your Ideal Customer Profile using a combination of firmographic and technographic filters, and then HubSpot tells you how many companies in the world match those criteria.
This last part is underappreciated. The ability to see the total addressable market in your CRM context, and compare it to how many you've already acquired, is a powerful strategic signal.
Consider a company targeting businesses that sell directly to consumers. The setup might look like this:
The technographic filter is crucial. It narrows the market to companies that are genuinely selling D2C and not just companies that say they do. If a business is running Shopify, they have a storefront, a customer base, and operational complexity that creates real buying triggers.
The resulting numbers might look something like this:
| Market | Total companies in market | Currently in CRM | CRM penetration |
|---|---|---|---|
| UK consumer goods (D2C) | ~80,000 | ~130 | ~0.16% |
| US consumer goods (D2C) | ~540,000 | ~70 | ~0.01% |
The takeaway here is strategic: the CRM already contains companies from both markets, which validates product-market fit, but penetration is extremely low. There are tens of thousands of qualified prospects not yet in the system. This is the gap that intelligent prospecting is designed to close.
Having a defined market is necessary but not sufficient. You need to know which companies in that market are actively in a buying mindset right now.
HubSpot's Buyer Intent feature does this by identifying companies that have visited your website based on IP resolution and company matching. Critically, you can filter this list to only show companies that:
This gives you a prioritized list of warm prospects: companies that fit your ICP and have already demonstrated curiosity about your product by visiting your site.
In practice, even a small number of intent-matched companies (say, 30–50) represents genuinely high-value outreach opportunities, far more so than thousands of cold contacts. These companies are already partway through a buying journey. Your job is to meet them where they are.
Website visitors who match your ICP are worth treating differently from cold prospects. The outreach message can acknowledge that you're reaching out because you noticed them in the space, without being creepy about it. The tone is warmer, the value proposition more targeted. Response rates for intent-matched outreach can be 3–5x higher than cold outbound.
Buyer Intent identifies companies visiting your site. Research Intent casts a wider net. It identifies companies in your target markets that are actively researching topics relevant to your product, even if they haven't visited your website yet.
This works through advertising pixel networks that track which content companies are consuming across the broader web. HubSpot categorizes this content and surfaces companies showing in-category research behavior.
For a ecommerce focused product, for example, Research Intent might surface companies actively reading about cart abandonment, checkout optimisation, subscription growth, or reducing return rates. All strong signals that they have a relevant pain point, even if they haven't discovered your solution yet.
A Research Intent list of 1,000+ companies is qualitatively different from a cold market list of 80,000. These are companies that have raised their hand by behavior, not just by fitting a profile. They belong at the top of any outreach sequence.
The most valuable segment is the intersection:
Running separate sequences for each tier with different messaging, cadences, and value propositions will outperform any single-sequence approach.
HubSpot's intent signals are company-level. They tell you which organizations are showing interest. But not who to contact within those organizations. You still need to identify the right people and obtain their contact information.
This is currently a manual enrichment step for most teams, using tools like:
The workflow looks like this:
This step is ripe for automation. Apollo and other tools have APIs and native integrations that can automate persona identification and contact enrichment for matched companies. The manual version works at lower volume; automation becomes essential once the pipeline scales.
Once you have qualified contacts in HubSpot from your intent-matched company list, the AI Prospecting Agent takes over the outreach personalization.
The agent (found under Prospecting Agent in HubSpot) works by:
The quality of the AI's output depends heavily on the selling profile you configure. This should include:
The more specific the selling profile, the better the output. Generic profiles produce generic emails.
A sensible rollout has two phases:
Phase 1: Supervised: Every AI-drafted email gets reviewed by a human before sending. This lets you evaluate quality, correct errors, and fine-tune the selling profile based on what looks good versus what needs work. Run this phase until you're consistently satisfied with the output, typically a few weeks of calibration and fine-tuning.
Phase 2: Autonomous: Once the agent is producing consistently high-quality emails, switch to autonomous mode. The agent researches, writes, and sends without requiring review. Volume scales dramatically while effort stays flat.
The supervised phase is not a bottleneck but an investment in calibration. Skipping it and going straight to autonomous often produces mediocre outreach at scale, which is worse than doing nothing.
Here's the full prospecting workflow assembled:
DEFINE → Target Markets (ICP + technographic filters)
SIGNAL → Buyer Intent (website visitors in target market + Research Intent (topic researchers in target market)
PRIORITIZE → Tier 1 (website visitors) → Tier 2 (researchers) → Tier 3/4 (cold)
ENRICH → Use Apollo/Lusha/ContactOut to find contact-level data
IMPORT → Add enriched contacts to HubSpot, linked to companies
PERSONALIZE → AI Prospecting Agent researches + drafts emails
REVIEW → Human review (supervised mode) or autonomous sending
OPTIMIZE → Track reply rates → refine selling profile → improve targeting
Skipping the technographic filter. Defining a market by industry and geography alone is too broad. Adding a technographic criterion, like Shopify usage, cuts the noise dramatically and ensures you're reaching companies with a specific operational context relevant to your product.
Treating all intent signals equally. A company that visited your pricing page three times this week is not the same as a company that read one blog post six months ago. Build your tiering logic around recency and depth of engagement.
Enriching contacts before defining intent tiers. Contact enrichment costs money and time. Do the intent filtering first, then only enrich the highest-priority tiers. Don't buy 10,000 contacts when 300 high-intent ones will outperform them.
Deploying the AI agent without calibration. The AI Prospecting Agent is powerful but not magic. It needs a strong selling profile, and it needs human review during the initial rollout phase to catch where it's going wrong.
Ignoring CRM penetration as a success metric. The Target Markets view gives you a denominator: total companies in your defined market. Track your CRM penetration over time. Growing from 0.01% to 0.5% penetration in a market of 500,000 companies is a different kind of progress than just counting meetings booked.
The full potential of this stack isn't just smarter prospecting. It's a closed loop. As the AI agent sends emails and you track responses, those engagement signals feed back into your intent data. Companies that reply become warmer signals for others in the same segment. Patterns emerge around which research topics predict conversion, which technographic combinations show up in your best customers, and which personas respond to which value propositions.
The manual enrichment step, currently the biggest friction point, will increasingly be automated through native integrations between HubSpot and enrichment providers. When that loop closes, the workflow becomes nearly fully automated from signal detection to first outreach, with human judgment focused on strategy and calibration rather than execution.
That's the trajectory: from spray-and-pray outbound to a precision signal machine where every email sent is justified by intent data and personalized by AI. The tools exist today. The teams that figure out how to combine them will have a structural advantage in pipeline generation.
This approach is most effective when your HubSpot instance is kept clean. Accurate company records, consistent contact association, and regular review of intent thresholds. The better the data hygiene, the more reliable the signals.