How to Spot Your Next Case Study Before Your Client Even Knows They're a Fit

Ever sat down to write a case study and drawn a blank on who to feature? Most teams rely on someone remembering a happy client months after the moment's passed, which means the best stories often go untold. A case study identification agent solves this by continuously scanning your company records for the signals that actually predict a good story - strong renewals, expansion revenue, high satisfaction scores, long tenure - and surfacing a shortlist before you go looking for one. This article covers what the agent does, where it fits in your existing process, and what to think through before you build it.

Example use cases this suits well:

- A customer success team wants a monthly shortlist of accounts hitting renewal or expansion milestones, rather than relying on a CSM to flag it manually when they happen to remember.

- A marketing team is planning a case study for a specific industry vertical and needs the agent to filter company records down to strong-performing accounts in that sector.

- An account manager closes a major upsell and wants to know immediately whether that account also has the satisfaction and tenure signals to make a credible public story, not just a big number.

The idea, in short: The agent runs on a schedule (or on trigger events like a renewal or NPS submission), checks company and deal records against a defined set of case study criteria, and produces a ranked shortlist with the specific signals that qualified each account - so a human decides who to approach, rather than the agent picking blind.

Top tip: score for "story," not just for size.

The biggest account isn't always the best case study - a smaller client with a dramatic before/after (steep ticket reduction, fast time-to-value, a clear turnaround) usually outperforms a large logo with an unremarkable journey. Weight the criteria towards contrast and change, not just contract value.

Who should own this: The customer marketing or content marketing lead, since they're the one who ultimately needs the shortlist to be usable - not just accurate. Customer success should be a required input into what "good" looks like (they know which accounts are genuinely happy versus quietly tolerating you), but ownership of the agent's criteria and output sits with marketing.

Considerations for the build:

- Decide which properties actually predict a good case study in your business - renewal rate, NPS/CSAT score, deal expansion, tenure, and support ticket trend are common starting points, but they need calibrating against case studies that have actually worked for you in the past.

- Set a minimum data threshold before a company is eligible for scoring - an account with three data points shouldn't rank the same as one with three years of history.

- Build in a mandatory human review step before any client is contacted. This agent should shortlist, not approach; letting it (or an over-eager team member) skip straight to outreach risks approaching a client who isn't actually willing, or who churned quietly after the last good quarter.

- Decide how the shortlist reaches the right person - a weekly digest, a Slack notification, or a property flag on the company record are all viable, but pick one and make it someone's job to act on it.

- Agree a re-scoring cadence. Case study fit changes - an account flagged six months ago may have since had a rocky renewal.

Common mistake to avoid: Treating the agent's shortlist as a finished decision rather than a starting point. Teams that skip the human sense-check end up either approaching a client who's about to churn (because the AI scored last quarter's data, not this quarter's mood) or missing the fact that a technically strong account is going through a leadership change and simply won't have bandwidth. The agent flags fit on paper; a person still needs to know the account.

Once it's running well, the sign that it's working isn't a bigger case study pipeline - it's fewer stalled conversations, because every client you approach was already a genuine fit before anyone picked up the phone. Pair it with the case study pipeline hack once a candidate says yes, and the whole journey from "who's a good fit" to "published story" runs on the same data.

Frequently Asked Questions

What HubSpot data does an agent like this actually need to work?

At minimum, company or deal properties covering renewal status, deal value/expansion, and some satisfaction signal such as NPS or CSAT. The more consistently your team logs this data, the more reliable the shortlist — an agent can't score what isn't in the record.

Does this replace the need for a customer success manager to nominate case studies?

No — it's meant to surface candidates a CSM might not think to flag, not replace their judgement. The agent should widen the pool a human then reviews, not make the final call.

What HubSpot tier or tools does this require?

This depends on which HubSpot AI/agent tooling and hub tier you're on. Flagging for your confirmation before publishing, since tier and pricing details shouldn't go out as unverified claims.

What's the biggest reason a shortlisted account turns out not to be a good fit after all?

Usually timing — the data reflects a strong quarter, but something's changed since (a champion has left, a renewal has gone quiet). This is exactly why a human review step matters before outreach.

Is this different from HubSpot's own AI-assisted case study creation tools?

Yes — HubSpot's native tools help you build and publish a case study once you know who it's about. This agent solves the step before that: working out who to ask in the first place.