List building before AI, list building after AI. Same logic, different scale.

I've spent close to fifteen years in B2B sales. I closed 600 self-sourced deals myself at Trustpilot. I ran a sales consultancy that worked with 100+ teams. And the most consistent pattern I've seen is this: top of funnel decides everything.

List building best practices existed long before AI. The logic hasn't changed. The leverage has. What used to be a junior rep's Monday-morning chore can now run as a discipline upstream of the whole team. The teams pulling ahead are the ones who realized that. The teams falling behind are running the same tools without the discipline underneath.

Here is what disciplined list building actually looks like in 2026, and what AI changes about who does the work. To make it concrete, I'll thread one example through the whole piece. A marketing services agency selling CGI, 3D animation, and technical visual content to B2B manufacturers. Their buyers are CMOs and brand leaders at companies that build complex physical products. We'll come back to them in every section.

01Bad top of funnel kills the rep two ways

9 out of 10 pipeline failures trace back to list quality, not outreach, timing, spam, or activity.

I have always said there are five reasons your pipeline isn't working. The five reasons are always the same.

1. List quality. Wrong companies, wrong contacts, or both.
2. Outreach. Bad cold call script, bad email copy, bad LinkedIn DM.
3. Timing. Calling during Christmas or summer holidays. Predictable. Should be planned for.
4. Spam. Your domains are burning, your numbers are flagged. Deliverability is broken.
5. Activity. Reps not picking up the phone. The only one fully in their control.

9 out of 10 times, it was a list issue.

The other four reasons get blamed first because they are easier to see. The cold call script is right there to review. The send time is right there to A/B test. The bounce rate is right there in the deliverability dashboard. Activity is right there in the CRM. List quality lives upstream of all of that, which is exactly why it gets diagnosed last. Teams spend months tuning the outreach copy of an unfixable list.

One note on Timing before we move on, because it has more depth than the list suggests. Timing has two layers.

Macro timing is calendar-driven. Christmas, summer holidays, fiscal year ends, major industry events, regulator deadlines. Shared across the whole market. Predictable. Easy to plan for. Most teams already do.

Micro timing is account-specific. Each company runs its own internal calendar, and the events on it move conversion far more than the macro layer does. The signals that actually matter look like this:

• A new VP of the function you sell into started this quarter. Their first 90 days are an open window.
• The CEO mentioned a "platform consolidation" initiative on the last earnings call.
• A job posting in Procurement just listed "vendor consolidation" as a responsibility.
• The annual report named a regional expansion in a geography you serve.
• The careers page is hiring for a new function that did not exist six months ago.

These are the "why this account, why now" moments. They are public, dated, and almost always visible to anyone who actually looks. The catch is that looking takes work. Tracking the micro layer manually across thousands of accounts is impossible. A rep can either follow a methodical process per account and burn hours per company, or skip it entirely and miss the signal. Most teams skip it. Or worse, they go down rabbit holes, spending forty minutes reading a 10-K and finding nothing actionable. Active recycling and signal-based outreach exist for exactly this reason. We will come back to who can do this work properly without burning the rep's week.

Sales is a funnel, not a cube.

Whatever quality you put in at the top is the ceiling for everything below it. Bad list, good outreach, you still lose. Good list, mediocre outreach, you compete.

If the top of your funnel is bad, your reps lose two ways.

Either they clean the list themselves, in which case they are not selling. They are spending hours per week reviewing accounts, deleting duplicates, checking whether contacts still work at the company, and trying to figure out which 20% of their list is worth their effort. Every hour spent on cleanup is an hour not spent on a call with a real buyer.

Or they don't clean the list, and they spend their week calling and emailing the wrong people. Bad-fit accounts at the top of the funnel do not just produce fewer wins. They produce slower discovery calls, stalled deals, weak forecasts, and reps who lose confidence in the list they were given.

Both failure modes look the same in the dashboard: low pipeline conversion, unpredictable months, frustrated reps. Disqualification is what a sales funnel does. Every stage qualifies in and disqualifies out. The point is not that reps should never disqualify. The point is where they should do it.

Top-of-funnel disqualification is the most time-consuming and the least valuable work in the funnel. It is hours per account, low signal, repeated across thousands of accounts. Bottom-of-funnel disqualification is minutes per account, high signal, against a handful of deals. Reps cost the same per hour at every stage. Where you put them decides whether you are paying premium labour for premium work or premium labour for cleanup.

Most teams I have seen do exactly this. Not because the leaders are wrong, but because there was no one else to do it. Until recently.

02Two motions, same problem: high-velocity vs enterprise

High-velocity and enterprise teams face the same upstream gap. The unit of work differs, the problem does not.

Two B2B teams can run wildly different motions and still face the exact same top-of-funnel problem.

A high-velocity mid-market team works hundreds of accounts per quarter, per ICP. One or two contacts per company. The bottleneck is volume and disqualification at scale. They cannot inspect every account by hand. They need an upstream process that ranks the list, surfaces fit, and discards the rest before any rep touches it.

An enterprise team works a few dozen named accounts per ICP. Twenty contacts per account, mapped across a buying committee. The bottleneck is depth, not volume. Every account is researched in detail. Every contact has a role on the committee. The same disqualification work still has to happen, but the unit of work is different. They are not picking accounts in or out as much as they are picking which committee members matter inside an account that was already in.

Same problem in both cases: bad top of funnel kills the rep. Different shape: volume-and-cherry-pick versus depth-and-committee. Most of the rest of this piece applies to both motions. I will flag where it does not.

03ICP is not industry, headcount, and country

A real ICP has more judgment criteria than filter criteria. If yours fits in three LinkedIn fields, it is too loose.

If your ICP definition fits in three LinkedIn filters, your list quality will be terrible.

Industry, headcount, and country are necessary, but they are nowhere near sufficient. Real ICP criteria are subjective. They require judgment about the company. That is the whole point.

Three real ICPs from three different B2B companies, side by side.

The marketing services agency selling to manufacturers has five criteria. Three of them require judgment. "In-house marketing leadership" cannot be filtered from a database. Someone has to look at the company and decide. "Manufactures a technical product that needs visualization" is the most important criterion of all, and the most judgment-heavy.

The transfer pricing data vendor has six criteria. Four require judgment. "Active cross-border intercompany flow subject to OECD" cannot be filtered from any database. Someone has to read the company's annual report, UK tax strategy PDF, or corporate governance disclosures and decide.

The customs compliance vendor has four criteria. Three require judgment. "Trader, importer, or exporter" sounds simple until you realize most companies do not advertise that distinction on their website. Someone has to read pages of the company's content and decide whether they are the buyer or the broker.

This is the work an Account Qualification agent is designed to do. It opens the company website, the careers page, the product pages, the regulatory disclosures, and decides on each criterion with evidence. ICP yes or no, with reasoning the rep can read if they want to.

The classic illustration of why filters alone fail is retail. Software companies that sell to retail will list themselves as "retail" on LinkedIn. Retailers will list themselves as "apparel" or "food" depending on what they sell. If you filter by industry on LinkedIn, you miss both ends.

Now the numbers. With a raw list pulled from a database and good qualification criteria, expect 30 to 40% of accounts to actually be ICP at best. For very niche products with strict criteria, expect 7 to 10%. Counterintuitively, the 7% list converts exponentially better than the 40% list. Tighter ICP, fewer accounts, much higher conversion.

The data you get out of a database is the floor of your list quality. The qualification criteria you apply on top is the ceiling. Both live in the Agent Training Center. That is where you write the criteria, define the verticals, and configure the personas the agents will use across every workflow.

Worked example — CGI agency ▾

Company: Marketing services agency selling CGI/visualization to B2B manufacturers.

1. Revenue above £1M — Filterable. Available in company databases.
2. Business activity in target verticals — Judgment. "Manufactures industrial products" is not a standard SIC code. Requires reading the company's website.
3. Geographic presence in US/CA/EU — Filterable. HQ location is in most databases.
4. In-house marketing leadership — Judgment. Need to check whether the company has a marketing team vs. outsourcing everything to an agency.
5. Manufactures technical product needing visualization — Judgment. The most important criterion and the most judgment-heavy. Requires looking at the product portfolio.

Result: 2 filterable, 3 judgment. More than half require judgment, so this ICP is tight enough. The judgment criteria are exactly what Account Qualification would evaluate.

04Personas are not job titles

The persona is the constant. The job title is the variable. Define the outcome, not the title.

The same problem repeats at the contact level, just less obviously.

Job titles are not unified across companies. The person responsible for sales enablement at Company A might be called "Product Marketing Manager" at Company B and "Sales Operations Lead" at Company C. The person who owns supply chain at one company is "Logistics Manager" at the next and "Head of Operations" at the one after that. Same persona, different titles.

A persona is the person responsible for an outcome, not the person with a specific title. That is the most important shift to make.

For the manufacturing example, the buyer is "the person in charge of brand and visual content at this manufacturer." At a $5B truck maker that might be a VP of Corporate Branding and Digital. At a glass manufacturer it might be a Head of Marketing who happens to own brand. At a medical device company it might be a Director of Product Marketing. Three companies, three titles, one persona.

The same shape repeats at the customs compliance vendor. The buyer is "the person who owns customs documentation." At a fashion retailer that is in Finance. At an industrial manufacturer that is in Trade Compliance. At a logistics-heavy importer that is in Supply Chain. Same outcome. Three different home departments. Three different titles. Boolean searches^* for "CFO" or "compliance manager" miss two of the three every time.

And the reverse trap is just as bad. Search LinkedIn for "CPO" and you get three completely different buyers. At a SaaS company it is the Chief Product Officer. At a fast-growing startup it is the Chief People Officer. At an industrial manufacturer it is the Chief Procurement Officer. Same three letters, three completely different people, three completely different sales conversations. The boolean search does not know the difference. A persona-aware agent does, because it reads the company and decides.

* A boolean search uses operators like AND, OR, and NOT to combine keywords. On LinkedIn this typically looks like ("VP of Sales" OR "Head of Sales") AND ("SaaS" OR "software"). Useful for narrowing or widening a list by title, but useless at telling whether a "CPO" means Product, People, or Procurement.

What good personas actually contain:

• The expertise and responsibility (who owns this outcome at this company)
• The pains, frustrations, and objections (different by industry and company size)
• The decision-making weight (decision maker, influencer, individual contributor, blocker)
• The DISC profile and communication style, where you can get them

This is exactly what the Contact Finder agent uses persona cards for. It does not just match titles. It reads About pages, LinkedIn profiles, and articles, and decides who actually owns the outcome at each company. When the right contacts are found, the Email and Phone Waterfall cascades through 20+ vendors to enrich them.

Good salespeople do this manually. They look at Sales Navigator, read About pages, study how the org is structured, and use judgment to decide who actually owns what. The work is real. It is also exactly the kind of work that scales when you stop relying on title strings and start interpreting context.

This is the structural reason data vendors like ZoomInfo, Cognism, and Apollo are not enough on their own. They give you fields. They do not give you context. Cleaning their output is exactly the top-of-funnel work you do not want your reps doing. The right question is not "how do we find more leads." It is "who or what should be doing the top-of-funnel disqualification, so that by the time a rep sees an account, the list is already clean." That is where list building stops being a manual process and starts being a designed one.

05Mid-market is dynamic. Enterprise is static.

Your TAM-to-CRM ratio decides which list-building motion comes first. Most teams guess this wrong.

Designed list building starts with the shape of your TAM.

If you target enterprise, your TAM is essentially fixed. A few hundred to a few thousand accounts globally. The list barely changes year to year. New entrants are rare. Acquisitions consolidate the list rather than grow it. The work is finding which accounts in that fixed TAM are ready right now.

If you target mid-market, your TAM is constantly being created and destroyed. New companies launch every week. Old ones die. Funding rounds, leadership changes, and product launches reshuffle the list. The work is permanent gap-filling alongside permanent recycling.

Either way, most teams treat list building as if there is some giant undiscovered market sitting outside their CRM. For 90% of established B2B companies, this is wrong. Your TAM has a size. That size is mostly set by the segment you target. Match it against what is in your CRM and the answer becomes obvious.

The marketing services agency in our running example sits in the mid-market band. B2B manufacturers with the budget and product complexity to need CGI work are not a long list. There are maybe a few thousand globally that fit. But recall from Section 03 that three of their five ICP criteria require judgment. That means you cannot bulk-pull prospects from a database. After five years of manual prospecting, the agency might have 750 companies in its CRM, roughly 25% of the true TAM. The job is not recycling what they already have. The job is finding the three-quarters they have never seen.

At the niche enterprise end the math is even more extreme. A vendor selling transfer pricing data to multinationals with €100M+ revenue, multi-entity structures, and active OECD exposure is working a TAM of maybe 3,000 globally. Most of that is sitting in any CRM that has been around a few years. The work there is almost entirely surfacing and prioritizing what they already have.

The first list-building mistake is misreading this ratio. Teams build elaborate gap-finding workflows when the answer was sitting in their CRM the whole time. Or they only mine their CRM when their TAM is dynamic and they should be doing both. Look at the size of your TAM relative to the size of your CRM. The ratio tells you which job comes first. The static-TAM team's job is mostly hygiene and recycling, which is what the Newbiz Recycling playbook is built for. The dynamic-TAM team's job is permanent gap-filling, which is what Newbiz Gap is built for.

Worked example — CGI agency ▾

Company: Marketing services agency selling CGI to B2B manufacturers.

1. TAM estimate: ~3,000 B2B manufacturers globally with the budget and product complexity to need CGI work.
2. CRM count: ~750 companies in CRM after 5 years of operation.
3. Ratio: 750 / 3,000 = ~25% penetration.

Why so low? The ratio reflects the ICP. In Section 03 we saw that 3 of 5 criteria require judgment: target vertical, in-house marketing leadership, and "manufactures a technical product needing visualization." None of those can be bulk-pulled from a database. Five years of manual prospecting surfaces maybe a quarter of the true TAM. Compare that to an employee-benefits vendor targeting companies with 1,000+ employees, where headcount alone is a database filter and the CRM fills itself. The more judgment-heavy the ICP, the lower the natural penetration.

Verdict: Below 40%. This agency has a real gap. The primary motion is Newbiz Gap — systematically find the ~2,250 companies that fit the ICP but are not yet in the CRM. That is exactly the work Newbiz Gap is built for, and exactly where Account Qualification pays for itself: it evaluates the three judgment criteria at scale instead of one-by-one.

06Your CRM is messy because B2B data is structurally messy

Your CRM is messy because B2B data arrives from too many sources that disagree with each other. That is structural, not a process failure.

Once you accept that most of your TAM is already in your CRM, you have to deal with the state your CRM is actually in. And the honest answer is: it is a mess.

This is not the fault of your reps. It is not the fault of your RevOps team. B2B data is structurally messy because it comes from too many sources that do not agree with each other.

Some of it came from event lists, where people self-entered their company name into a badge form. Some came from data vendors who normalized differently. Some came from reps typing manually into the CRM late at night. Mix all that together over a few years and you get duplicates, mismatched names, accounts marked "open" that should have been closed-lost three years ago, and a long tail of companies someone added because "I think I can close them."

If you work with salespeople who prospect, your CRM will have duplicates. That is a feature of how the data flows in, not a process failure. Get comfortable with it.

There is also a quieter form of decay. Roughly 3 to 4% of contacts in any B2B CRM lose accuracy every month because people change jobs. Over a year, that is a third of your contact base. By year two, more than half is wrong.

Imagine a vendor selling customs and trade compliance software with 4,500 customers across the Nordics, all of them parent-child structures. Same brand operating as separate legal entities in Sweden, Norway, Denmark, Finland. The CRM cannot represent this cleanly. Some entries are merged, some are duplicated, some are linked, some are not. This is not a bug. This is what B2B data looks like at scale.

The cleanup is two jobs, not one.

The first is merging duplicates and tidying account hygiene. The second is harder and more important: re-qualify everything against your current ICP and current personas. A company that was ICP three years ago might not be today. A contact who was your champion last year might be at a different company now. Run Account Qualification across your CRM accounts to re-check ICP fit against your current criteria. Run Contact Qualification across your CRM contacts to check whether they still work at the company and still match your buyer personas. Until the re-qualification happens, the volume number in your CRM is fiction.

07Where lists come from depends on where your customers self-list

Companies self-list wherever they need to be found. Understand the five reasons and you know where to look.

Once the existing CRM is clean, the gap remains: the companies that should be in your CRM but are not. The next question is where to find them.

Companies do not pick categories at random. They categorize themselves wherever they need to be found. If you understand why they need to be found, you know where to look.

There are five reasons a company will self-list somewhere public.

They want to find customers. This is the biggest one. Local services like restaurants, clinics, carpenters, and dentists show up on Google Maps because that is where their customers look. Software companies show up on LinkedIn because that is where their customers and talent live. B2B businesses with a marketing budget show up on trade-show exhibitor lists. Products sold through a partner network, like cars, furniture, bicycles, or IoT hardware, show up in distributor and partner directories.

They want to find talent. Job boards and LinkedIn. The more niche the industry, the more there is a vertical-specific job board where the same companies post over and over.

They are regulated. Fintech, real estate, gambling, law firms, accounting firms. Anything where a regulator licenses operators. The regulator publishes the list. So does the local company registry.

They want to lobby. International, national, and regional industry associations publish their members. If your buyers care enough about an issue to pay association dues, they care enough to get on a sales call.

They are ranked. Industry rankings, "top X" lists, category awards. Anyone proud enough to be on a list is generally pre-qualified on at least one dimension.

Once you know the source, the operational question is buy or scrape. You can buy slices of these databases from Similarweb, BuiltWith, or LinkedIn through tools like Snov.io. You can scrape directly using Fiverr, webrobots.io, or snov.io. Either path gives you a raw list. Neither gives you a clean one.

For our marketing-services-for-manufacturing example, the right starting points are trade-show exhibitor lists from industrial expos plus distributor and partner directories. Both are dense with companies that build technical products and care about visual content. This is the kind of work the Newbiz Gap playbook is built for.

08Three sources of pipeline: Gap, Recycling, Expansion

Gap, Recycling, and Expansion are three different jobs running on different cadences. Blend them and you do all three badly.

Once you know how big your TAM is and how much of it sits in your CRM, you have three sources of pipeline, not two. Most teams need all three. The mix depends on segment.

Newbiz Gap is for companies not yet in your CRM. The more dynamic your TAM, the more permanent this source needs to be. Companies launch, companies die, new buyers emerge. Even at the enterprise level, where TAM is essentially static, an occasional gap pass catches the new entrants.

Newbiz Recycling is mining your existing CRM. There are three flavors worth naming, because they look different in practice.

The first is active recycling. Deals that cooled down and are now ready to be picked back up. The trigger might be a leadership change, a new round of funding, a product launch, a hiring spree, or just twelve months of silence. These are warm. This is the micro-timing layer from section 01, applied systematically. This is where Account Research and signal-based outreach earn their keep. Schedule research to run on a cadence and let the system surface the accounts where something has changed.

The second is legacy backlog. Companies that have been sitting in the CRM untouched for years. Nobody owns them. The data is stale. They need re-qualification and re-research before any rep should touch them. These are not warm, but they are cheap to surface.

The third is champion monitoring. The cleanest version of this I have seen comes from a customs compliance software vendor. Their buyers are finance leaders. When those finance leaders change jobs, the first thing they often do is call the vendor to get the product installed at the new company. That single behavior, monitored systematically, is the entire ROI of watching people instead of just companies. The Champion Monitoring playbook automates this.

Account Expansion is the third source, and the one most teams underuse. List building inside an existing customer is still list building. Big customers have multiple regions, multiple business units, multiple buying committees that do not talk to each other. The expansion gap inside one enterprise customer is often bigger than landing five new mid-market logos.

A corporate travel and expense management platform is the cleanest illustration. They land on a single business unit, often through a finance leader. From there, every region, every subsidiary, every newly acquired company is a separate expansion list. Same product, different buying committee. The buying-committee map you used to land them is the same map you use to grow them, just applied one business unit at a time.

The trap is treating these as one. Active recycling is sniper work. Legacy backlog is volume work. Champion monitoring is signal work. Expansion is map work. They run on different cadences, with different research depth, against different signals. If you blend them, you do all four badly.

If you run them well, the same company comes through your pipeline three or four times before it ever reaches the ceiling on your relationship with them. Once as a gap-fill. Once as a cooled-down opportunity. Once as a re-qualified backlog account. Once or several times as expansion within the same logo.

Worked example — CGI agency ▾

From Section 02: Mid-market motion (hundreds of accounts per rep).

From Section 05: TAM penetration ~25%. TAM is mostly static (B2B manufacturers are a stable population), but three-quarters of it is unknown to the CRM because the judgment-heavy ICP blocked bulk discovery.

Decision: TAM penetration below 40% + static TAM → start with Newbiz Gap. Use Account Qualification to systematically evaluate the ~2,250 companies not yet in the CRM against the three judgment criteria. Run a secondary Newbiz Recycling pass on the 750 accounts already inside to re-qualify stale records.

Champion Monitoring is a natural second motion here — when a marketing leader at one of the qualified accounts changes jobs, that is an expansion signal worth tracking.

09What this looks like in practice: a plateaued enterprise team

A plateaued enterprise team almost always has the problem inside the CRM, not outside it.

Let me walk you through a real shape I have seen at multiple companies. Anonymized, but the pattern is consistent enough that it is worth describing in detail.

A B2B vendor in the market for five years. Enterprise focus. TAM around 2,000 large-corporate accounts with 1,000+ employees. Average ACV of €100K+. Six SDRs, three AEs, one Sales Ops Manager. About 400 customers. And for the last twelve months, sales had plateaued. Same team, same TAM, same product. Pipeline coverage was technically there. Conversion had stalled.

The instinct in that situation is always to look outside. New markets. New verticals. More leads at the top of the funnel. That instinct is almost always wrong for a mature enterprise team.

Look at the numbers. TAM of 2,000. CRM has 1,700 accounts. That is 85% TAM penetration. There is no gap to fill. The problem is not outside the CRM. It is inside it.

What ABS is, briefly

Before I walk through the diagnosis, one piece of vocabulary. ABS stands for Account-Based Sales. Every account in your CRM sits in exactly one of four states at any given moment. The state is determined by hard CRM criteria, not opinion. It tells the team what work belongs on that account, who should be doing it, and what counts as progress.

An account moves through these stages over time. A Dormant account becomes Working when a rep starts outreach. A Working account becomes Opportunity when a meeting happens and a sales cycle opens. An Opportunity becomes Customer on close, or returns to Dormant on loss. A Customer can spawn a new Working or Opportunity inside the same logo (expansion).

The point of putting every account in exactly one stage is to make the right work visible. Dormant accounts need re-qualification and recycling. Working accounts need outreach quality and timing. Opportunities need sales execution. Customers need expansion mapping. Four different jobs. Without the stages, those four jobs blur into one undifferentiated "pipeline."

The diagnosis

Three things, all of them upstream of the SDRs.

First, there were no ABS stages on accounts. Dormant, Working, Opportunity, Customer, that breakdown did not exist. Without those stages, there was no way to distinguish a cold dormant account from an active working account from a closed-lost one that had cooled. Reps treated all 1,700 accounts as equally workable. Which meant they treated them as equally unworkable.

Second, there was no reverse-engineered data on the 400 closed-won customers. Nobody had asked the question: what data points actually predicted that these accounts would close? Not headcount and industry, those filtered too loose. Specific signals. Did they use a particular technology? Did they have a specific organizational structure? Did they show specific hiring patterns in the year before they bought? The answers existed in the closed-won data, but nobody had pulled them out.

Third, the SDRs were doing manual research that should have sat upstream of them. Hours per week, per rep, on list cleaning. Hours per week, per rep, on identifying which accounts in the CRM were worth touching this quarter. Hours per week, per rep, on finding contacts and verifying they still worked at the company. The expensive misallocation from the previous section, running at scale across six SDRs.

The intervention

Three matched changes.

ABS stages went live on all 1,700 accounts. Suddenly the team could see that 60% were Dormant, 25% Working, 8% Opportunity, 7% Customer. The dormant pool was the prize.

The 400 closed-won customers got reverse-engineered. Account Research ran across them, looking for patterns. The signals that came out were specific. "Has internal compliance leadership," "operates across more than five countries," "uses a particular ERP," "hired in a specific function in the last six months." Those signals became enrichable fields in the CRM.

Then Account Qualification ran across the 1,300 dormant accounts using the reverse-engineered criteria. Not against headcount and industry. Against the actual signals that predicted conversion. The output was a prioritized list of dormant lookalikes: accounts that resembled closed-won customers, ranked by how many predictive signals matched.

Account Research layered on top of that, running on a cadence and surfacing the dormant accounts where something had changed: a leadership move, a funding event, a product launch, a hiring pattern. Newbiz Recycling became signal-driven.

The SDRs stopped cleaning lists. They started each week with a prioritized queue: dormant lookalikes plus active-recycling triggers. The qualification work happened upstream, against criteria the team had actually defined, at a scale the SDRs could never have done manually.

Same team. Same TAM. Same product. Different list-building discipline. The plateau was not a market problem. It was an unsystematized list-building discipline hiding twelve months of dormant-account opportunity.

If you want to see what this looks like in real customer environments, the customer stories from Luzmo and ARIS walk through similar shapes.

/The discipline is old. The leverage is new.

None of the disciplines I have walked through in this piece are new. They predate AI by a decade or more. The good salespeople I worked with at Trustpilot, in my consultancy, and now at Evergrowth all did some version of this manually. They mined CRMs that nobody else thought to mine. They wrote ICP definitions that took judgment, not filters. They built persona maps that took reading, not boolean searches. They disqualified accounts before anyone else got near them.

What is new is that the work is finally separable from the rep. The Account Qualification, Account Research, Contact Finder, and Contact Qualification agents can all run before a rep touches the list, against criteria you actually defined, at a scale that was not possible five years ago. Same logic. Different scale. That is what we built Evergrowth to do.