6sense Tells You Who. Demandbase Tells You When. Neither Tells You What to Say.

Every ABM team I talk to has the same stack. 6sense or Bombora for intent data. Demandbase or Terminus for account orchestration. Salesforce. Marketo or HubSpot. A custom tier model. A weekly account review where someone with the title "ABM Manager" reads off which accounts moved to "researching" and which fell back to "aware."

And every one of those teams has the same problem.

The platform tells them which accounts are heating up. It tells them the right week to act. It does not tell them what to say.

So the SDR writes one good email for the Tier 1 account, four templated emails for the Tier 2 accounts, and nothing at all for the Tier 3s. Multiply by 200 accounts in market this week, two SDRs on the team, and the gap between what the platform identifies and what actually gets shipped is bigger than most VPs realize.

That gap is where ABM execution goes to die. I've argued before that ABM tiering was a resource constraint dressed up as a strategy. This is what the constraint looks like up close.

The Tier 1, Tier 2, Tier 3 reality

I'm going to describe an ABM motion you'll recognize.

Tier	Account count (typical)	What each account gets	Who shows up
Tier 1	50	Custom research brief, personalized first-touch email, named buying committee, custom demo, personalized landing page	SDR + AE + Marketing + SE
Tier 2	250	Templated email with company name dropped in, generic LinkedIn message, generic one-pager	SDR only
Tier 3	1,000+	Quarterly newsletter, event invite, paid retargeting impression	Marketing automation only

How Tier 1 accounts get treated

Each Tier 1 account gets a real research brief: who the buying committee is, what the last earnings call mentioned, who left for a competitor recently, which integrations they shipped this quarter. The SDR writes a first-touch email that names three specific signals. The AE gets looped in for week two. Marketing builds a personalized landing page. Sales engineering preps a custom demo. The orchestration is real.

Why Tier 2 accounts get templated emails

Each Tier 2 account gets a templated email with the company name dropped in. Maybe a templated LinkedIn message. Maybe a generic value-prop one-pager. The SDR meant to do more, but the math didn't work. Twenty-five Tier 2 accounts per SDR per week, ten minutes of research each, that's already a full day before a single message ships. So the SDR ships the template.

What happens to Tier 3 accounts

The 1,000+ accounts the platform identified as in-market but didn't promote to a higher tier. These accounts get a quarterly newsletter, an event invite, maybe a paid retargeting impression. No outbound. No research. No conversation.

If your team is being honest, this is the picture. Some VPs will push back and say "we research every Tier 2 account too." If you do, you're either running a smaller book than you think or your team is burning out.

The dirty secret of ABM execution is that Tier 2 gets templated emails because no one has time to research them. Tier 3 gets nothing because no one is even trying. This is the same gap I described in Marketing has been selling ABM for 10 years. AI agents just made it real.

What ABM platforms actually do (and what they don't)

To be honest about what the intent and orchestration stack is and isn't:

The identification layer

6sense, Bombora, ZoomInfo Intent. These tell you which accounts are showing buying signals. They are very good at this. The data is real. The signals matter.

The orchestration layer

Demandbase, Terminus, RollWorks. These tell you when to act and across which channels. Account orchestration is a hard problem and these platforms solve real chunks of it.

The CRM layer

Salesforce, HubSpot. These hold the records and route the work.

The campaign layer

Marketo, Pardot, HubSpot Marketing. These run the campaigns.

The execution gap

None of these platforms writes the email. None of them does the prospect research that makes an email worth reading. They identify, route, send. The substantive content work, the part where someone reads the email and thinks "this is about me, not about everyone in my industry segment," is human work. It is the bottleneck.

Most ABM stacks have spent six figures on the identification and orchestration layers. The execution layer is still one SDR, one keyboard, ten minutes per account.

Why your SDR team can't close the gap

Walk over to the broader sales org and you see the same bottleneck from a different angle.

A full-time SDR costs $70,000 to $90,000 all in. They spend 60 to 70 percent of their time on research. That is not a productivity problem. That is what good outbound requires. I broke this down in more detail in Your SDR is spending 60 percent of their time on research.

The output ceiling

The output ceiling is 8 to 12 personalized emails per day. That math does not change with better tools. It changes with the prospect's name, their company, their role, the recent move that gives the email a reason to exist. Twenty to forty minutes per account, depending on how deep the SDR goes.

Eight to twelve per day, four days a week, two SDRs on a team. That's 64 to 96 deeply researched emails per week for the team. Compare that to the 250 Tier 2 accounts you identified this week alone.

The math doesn't work. It never works. And the gap between the math and the goal is exactly what gets templated.

Why hiring more SDRs doesn't fix it

This is the same gap the ABM platform creates and the same gap your SDR team can't close. It is not an ABM problem or a sales problem. It is a research problem that has nothing to do with intent data or orchestration software. You can hire your way to a bigger team and the Tier 2 problem just gets pushed down to Tier 3.

If you want to see the math on this, our ROI calculator compares SDR hire cost against AI-augmented research, including the soft costs (manager time, ramp, attrition) most pricing pages skip.

What "prospect research" actually means

When I say "research" I do not mean "look up the company's website."

For one mid-market SaaS prospect, real research is this list.

1. The CEO's most recent earnings call or board update. What did they tell investors about growth, hiring, and product priorities for the next two quarters?
2. The current org chart on LinkedIn. Who runs the function you sell into? Did they join recently? Who left? What did the person who left say about the org on the way out?
3. The hiring page. How many open SDRs, AEs, RevOps roles? Hiring a new VP of Sales or VP of Marketing is a signal that the function is being rebuilt. So is hiring an SDR Manager from a specific competitor.
4. The product. Did they ship a new integration in the last quarter? Did they launch a new tier? Did they change pricing? Each of these is a buying signal in your direction or your competitor's.
5. The competitor stack. What categories do they buy from? What did they recently rip out and replace?
6. The named pain. After 20 minutes of the above, you have a working hypothesis about what they're trying to fix this quarter. That hypothesis is the email.

Twenty to forty minutes per account, every account, every week. That is what the Tier 1 account gets and the Tier 2 account does not. The same logic applies even before you tier: this is the same kind of research that goes into defining your ICP when you're building the program from scratch.

You can hire more SDRs. The Tier 2 problem just gets pushed out to Tier 3. You can buy better intent data. The execution math stays the same. You can write better templates. The reply rate for templated email is what the reply rate for templated email has always been. (The difference between mail-merge and real personalization is the subject of its own post: AI personalization vs mail merge.)

The bottleneck is the research, and the research is the work the platform doesn't do.

How to close the ABM execution gap with AI

I'm not going to pretend AI solves every part of this. The contrarian version of "AI replaces SDRs" misreads what SDRs are good at. SDRs are good at the prospect-judgment work. They are good at reading a reply and deciding whether to push, soften, or back off. They are good at meeting prep and the messy work of getting an actual deal done.

What AI is actually good at

The work AI is actually good at is the part that kills pipeline quality. The 20 minute deep research on each account. The first draft of a personal-feeling email that names three specific signals. The cited stat that gives the email a reason to exist. The fact-check on whether a claim about the prospect's company is actually true.

That work, done at the pace of "every account, every week, including the Tier 2 and Tier 3 ones," changes what your sales team can ship. How AI outbound actually works walks through the full pipeline.

How the production line runs

Some specifics on what closing the gap looks like in practice, drawn from how we run our own program.

Per account. The engine does a 20 to 40 minute prospect research pass. Reads recent funding, hiring, product launches, integration releases, executive moves. Cross-checks against the named ICP signals. Notes the specific reason this account is in market right now. Account scoring sorts the in-market accounts by fit and signal strength.

Per prospect on the buying committee. The engine produces a personal first-touch. Not a template with a name field. A real email that names the prospect's specific situation. We run this through about ten guardrails inside our brand voice layer. No banned words. Cited stats only from a whitelist. The email reads like a person wrote it because the research is good enough that a person could have.

Per week. The team ships 150 to 200 of these. Across email and LinkedIn. Monday through Thursday for email, Monday through Friday for LinkedIn. The number is not 1,000. It is 150 to 200. The number is bounded by sending capacity and quality, not by what a tool could theoretically generate.

The result is that every account gets Tier 1 treatment. The Tier 1 accounts stay Tier 1. The Tier 2 accounts get a real first touch instead of a template. The Tier 3 accounts get a real first touch instead of a quarterly newsletter.

That's what closing the gap looks like. The full ABM execution service page walks through how this slots in alongside your existing 6sense or Demandbase rollout. We ran this on ourselves first, which I wrote about in We became our own first client.

You don't have a pipeline problem. You have a research problem.

This is the line I'd ask any VP of Marketing or VP of Sales reading this to sit with for a minute.

Your pipeline is what comes out the end of the funnel. If you are short on pipeline, the instinct is to add to the top. More accounts in market. More SDRs. More tools.

The funnel is not the problem. The work that happens before the email ships is the problem. The work the platform names "Tier 1 treatment" and your team can only afford to give to the top 50 accounts.

6sense tells you who. Demandbase tells you when. Neither tells you what to say. The "what to say" is the research, and the research is what you've quietly been rationing for the last two years because the math never worked.

If your team is at 8 to 12 personalized emails per SDR per day, you don't have a pipeline problem. You have a research problem.

If the only accounts that get real research are the top 50, and the next 250 get templates, you don't have a Tier 2 conversion problem. You have a research problem.

If your intent data is good and your reply rates are not, you don't have a copy problem. You have a research problem.

The shift is from "more accounts at the top" to "more research per account." Same accounts. More work per account. Pipeline math changes. The free Kill Your ABM Tiers playbook walks through how to run this shift inside a team that already has 6sense or Demandbase deployed.

Frequently asked questions

What do 6sense and Demandbase actually do for outbound execution?

6sense and Bombora identify which accounts are showing buying signals. Demandbase, Terminus, and RollWorks orchestrate when to act and across what channels. These platforms identify, route, and send. They do not do the per-account research that makes an email worth reading, and they do not write the email. The execution layer of every ABM stack is still one SDR at a keyboard with about ten minutes per account.

Why do Tier 2 accounts in ABM get templated emails?

Math. A typical ABM book has 50 Tier 1 accounts, 250 Tier 2 accounts, and 1,000+ Tier 3 accounts. A full-time SDR can produce 8 to 12 personalized emails per day because real per-prospect research takes 20 to 40 minutes. Two SDRs working four days a week ship 64 to 96 researched emails per week for the team. That math cannot cover 250 Tier 2 accounts. So Tier 2 gets the template and Tier 3 gets nothing.

How much time does an SDR actually spend on research?

60 to 70 percent of their week. That is not a productivity problem. It is what real outbound requires. The output ceiling of a full-time SDR is 8 to 12 deeply personalized emails per day, and that ceiling does not change with better email-writing tools. The bottleneck is upstream of the email.

What does real prospect research include?

For a mid-market SaaS prospect: the CEO's most recent earnings call or board update, the current org chart on LinkedIn (who runs the function, who joined, who left), the hiring page (open SDR, AE, RevOps roles, recent leadership hires), the product (new integrations, new tiers, pricing changes), and the competitor stack. After 20 minutes of that, you have a working hypothesis about what the company is trying to fix this quarter. That hypothesis is the email.

Can AI actually replace SDR research without losing quality?

Yes for the research and first-draft work. No for the prospect-judgment work. SDRs are good at reading a reply and deciding whether to push, soften, or back off, and at the messy work of getting a deal done. AI is good at the 20 minute deep research pass on every account, the cited-stat fact check, and the first draft of a personal email that names three specific signals. The right split is AI on the research bottleneck and SDRs on the conversation.

How does Agentic Demand close the ABM execution gap?

Per account, a 20 to 40 minute research pass on funding, hiring, product launches, integration releases, and executive moves, cross-checked against named ICP signals. Per prospect on the buying committee, a personal first-touch email that names the prospect's specific situation and passes about ten guardrails (no banned words, cited stats only from a whitelist). Per week, 150 to 200 of these across email and LinkedIn. The result is that every account gets Tier 1 treatment, including the Tier 2 and Tier 3 accounts the SDR team never gets to.

Is this a replacement for 6sense or Demandbase?

No. It sits next to them. 6sense and Demandbase keep doing what they do well (identification and orchestration). The execution layer plugs in alongside. If you're already paying for an intent or ABM platform, the research-and-email work is what's missing, not the platform itself.

See it on one of your accounts

If you want to see what Tier 1 research applied to a Tier 2 account looks like, the easiest way is to look at one of yours.

Pick an account in your ABM list that's in market this quarter and that your team hasn't given a personal email to yet. Send me the company name. I'll do the 30 minute research pass and show you what the engine produces. No charge, no obligation, no pitch deck.

Our goal at steady state is 5 to 15 meetings per month for a client. The reason that range is wide is that it depends on your ICP fit, your sender domains, and how aggressive you want to be on volume. We don't promise a number. We show you the output and let the output do the convincing.

Book a 30 minute walkthrough and I'll have an account brief ready when we meet. If you'd rather read first, the AI prospecting guide covers the methodology in more depth.

Related reading: ABM tiering was a resource constraint, not a strategy • Marketing has been selling ABM for 10 years • Your SDR is spending 60 percent of their time on research • How to define your ICP for outbound • How AI outbound actually works • AI personalization vs mail merge