Skip to content
Evergrowth Documentation

Account qualification

Account qualification is the gate that answers one question about a company before your team spends any time on it: is this a real fit? It is the first decision Evergrowth makes about an account, and everything downstream depends on getting it right.

What it is

Section titled “What it is”

Qualification is a structural-fit check. Your AI colleagues research a company against the criteria that define your ideal customer, then return a clear verdict and a written reason for it. The verdict is one of three: a fit, not a fit, or inconclusive when the public evidence simply was not enough to decide either way. The reason is the short explanation the agent records alongside the verdict - what it found, and why it landed where it did - so a “no” or an “inconclusive” is never a black box your reps have to argue with.

The third state matters. Qualification does not force a guess: an account it cannot confidently place stays flagged as inconclusive rather than being waved through as a fit or wrongly rejected. That keeps the “yes” list clean - everything in it cleared the bar on real evidence - and turns the inconclusive pile into a small, honest review queue instead of noise buried in either extreme.

It is deliberately the first thing that happens. Research, signals, persona matching, and outreach all build on top of a qualified account. If qualification is wrong, every later step is wasted effort aimed at the wrong company. Qualifying first is how Evergrowth keeps your reps working only the accounts worth working.

The criteria behind the verdict

Section titled “The criteria behind the verdict”

The fit check runs against the criteria that define each of your target segments. These are the conditions a company has to meet to count as one of yours: operating in the right region, being the right kind of organization, the handful of attributes where, if the answer is no, the account simply is not for you and no amount of good timing changes that. Keep this set tight and genuinely meaningful. The discipline of a focused criteria list is what keeps qualification honest - the moment loose, “nice to have” conditions creep in, you start rejecting good accounts and the gate loses its meaning. You define these criteria once per segment in the Agent Training Center, and the qualification check applies them to every account it researches.

A second, separate kind of criterion is not part of the gate at all: research criteria are additional facts you want the workspace to gather about accounts that have already qualified - context that feeds the deeper account research, account plans, and outreach built on top of a fit. These never decide whether an account passes; they enrich the picture once it has. Keeping the two apart is deliberate: the fit decision stays clean and explainable, while the richer context is collected only where it is worth collecting - on accounts you have already confirmed are worth the effort.

Eva term mapping: the gate criteria are the customer’s Qualification criteria (defined for the Qualification agent); the second, non-deciding kind are the Research criteria (defined for the Research agent - “additional data to be collected about accounts that pass qualification”). When a customer asks “why didn’t this account qualify,” the answer is always a failed qualification criterion, never a research criterion. The ICP verdict surfaces as three states - “Yes” (a fit), “No” (not a fit), and “Inconclusive” (not enough public evidence to decide); “Inconclusive” is a genuine third outcome, not a low score. The recorded explanation behind a verdict is the “reason.” There is no per-criterion weight or scoring layer in qualification - the account-level “Score” customers see is a separate signal-strength percentage from the Research agents, not a sum of criteria.

Why qualifying first protects rep time

Section titled “Why qualifying first protects rep time”

Reps are expensive and their attention is finite. Without a fit gate, that attention leaks across every account a list or an integration happens to surface, and most of it lands on companies that were never going to buy. A clear fit verdict at the front of the process is the cheapest filter you have: it spends a few research credits to save hours of human time chasing the wrong logos.

It also keeps the whole chain of agents pointed at the right targets. Account research only runs deep on companies that passed. Signal monitoring only watches accounts that matter. Outreach is only ever written for a genuine fit. The fit gate is the multiplier that makes every downstream layer worth running - and the reason qualification has to be right before anything else fires.

There is a compliance dimension too. Qualification is a company-level check that uses only public information, and it has to clear before any contact-level work begins. Personal data is never touched until the company itself has qualified as a fit. See Data, privacy & GDPR for how this gate is enforced.

Why verifying what a company actually does matters

Section titled “Why verifying what a company actually does matters”

The most common way qualification goes wrong is subtle: an account passes all the generic checks - right size, right geography, right region - and gets sorted into a target segment whose business it does not actually match. A logistics firm slips into your healthcare segment because it ticks every box except the one that matters most: what the company genuinely does.

That is why qualification always verifies a company’s core business activity, not just its surface attributes. Confirming what a company actually does is the check that ties an account to the right part of your ecosystem and verticals. Skip it and you get accounts that look qualified on paper but are aimed at the wrong value proposition, which then poisons the research and outreach built on top of them. Verifying the business itself is what makes “qualified” mean qualified-for-this-segment, not just qualified-in-general.

When a human knows better than the agent

Section titled “When a human knows better than the agent”

The agent’s verdict is a strong default, not the final word. A rep or sales leader can override an account’s ICP status by hand - mark a “no” as a “yes” against the right segment, or push a borderline account the other way - whenever they have context the public record does not carry. A manual change is recorded as exactly that: the account is flagged as manually set, and its reason captures who changed it, from what to what, and when, so the human decision is auditable rather than silently overwriting the agent’s work.

A separate, deliberate choice is whether to lock an account’s qualification. A locked account is left alone by every later qualification run - agents and workflows skip it and its status stays exactly as it is. Locking is how you protect a hard-won human judgment from being re-decided by the next automated pass, or freeze an account you have settled on. An unlocked account, by contrast, can be re-qualified as the company’s evidence changes, and a fresh agent verdict clears the manual-override flag again. The distinction is worth keeping straight: an override changes what the status is; a lock changes whether it can be changed again automatically.

Finding a domain before qualifying

Section titled “Finding a domain before qualifying”

Qualification is only as good as the company it can research, and research starts from a domain. An account that arrives without one - a bare company name from a list or an event - has nothing for the agent to investigate, so it would stall before it began. To close that gap, qualification can be told to first send any domain-less account through the Domain Finder, which locates the company’s website, before the qualification work runs. It is an optional pre-step you turn on when you are qualifying a batch that may be missing domains; accounts that already have one are unaffected. This keeps a messy import from quietly producing a pile of un-researched accounts. See how to run qualification for how to enable it.

In practice

Section titled “In practice”

A team selling compliance software to financial institutions might define its qualification criteria as: the company is a regulated financial institution (get this wrong and nothing else matters), and it operates in a covered jurisdiction. An account missing either of those never enters the pipeline. For the accounts that pass, the team adds research criteria for things like multi-entity structure or an existing compliance team - context the workspace then gathers to make later research and outreach sharper, without ever changing the fit decision itself.

The result is a pipeline where every account is a real fit, with the deeper context already gathered where it counts - before a single rep spends a minute on outreach.

Section titled “Related”