The true value of MarTech in 2026 is not only the data a tool can output, but what you can actually do with the data in practice.

Imagine hopping on a Zoom call and being prepared with your account's buying propensity with human-readable reasoning explaining it. You know the name of the engineer who configured SSO last week, the VP who viewed pricing twice, and someone from security who joined the workspace on Monday. You read it in about four seconds and lead the call with the security playbook.

Trailspark ingests your first party signals from your entire ecosystem, from marketing to product events, evaluates your accounts and writes the results into your CRM as native properties. But the reasoning data isn't just a "nice to know"... it's designed to help sales reps take the right next step when working their accounts.

Set it up once, then it pays out throughout the entire workflow: the ten minutes before a call, the call itself, and the follow-up. Here's how to set it up.

" There's no shortage of data in modern B2B tech stacks. The job is making it easy for your reps to find when they need it, and know how to interpret it. "

Step 1: Send Trailspark your signals

Nothing gets evaluated until Trailspark can see what people are doing. Anything that can make an HTTP request can send signals, and most teams pull from some mix of:

Product usage - The timeline of events straight from your app. Every time a feature is used, onboarding milestones, collaboration, API usage thresholds, subscription changes. In a PLG motion this is the good stuff, and it's the source most scoring tools never do anything meaningful with
Marketing automation - Marketo, HubSpot workflows, or any platform with outbound webhooks. Form fills, email engagement, webinar attendance, and page visits
Your CRM - HubSpot or Salesforce contact and account properties for added context about the account's current lifecycle stage or motion

You don't have to standardize any of it first. Each source sends data in its own shape, and Trailspark's mapping layer normalizes everything into one unified signal stream. Your product team's events and your marketing team's webhooks feed the same evaluation without anyone rewriting their instrumentation to match a spec.

The Syncs screen in Trailspark, described as defining what to sync from your warehouse and how often. Four syncs are listed and all show a green Synced status: Product Events from PostHog running hourly, Product Events for Feature Adoption running daily with 20 rows, Accounts running daily with 5,455 rows, and Users running daily with 5,602 rows. Each row has toggles and controls for Run now, History, and Edit. — Sources arrive on their own schedules and in their own shapes. Product events, accounts, and users all land in the same signal stream.

Identity resolution then runs on its own, tying every fragment to the right person and account. That includes product users who never spoke to sales, and anonymous visitors whose earlier activity attaches retroactively once they identify themselves.

Enrichment is optional and works behind the scenes. Connect your own Clay or Reo.dev account and it fills in company details your CRM doesn't cover. It only adds detail to accounts you already track, and your CRM's values always win a conflict.

Step 2: Describe the accounts you want to hear about

This step is important; it shapes everything downstream, so give it some real thought.

In Trailspark, an ideal-customer definition is a plain-language description where you describe what a good account looks like in plain English, the way you'd explain it to someone joining your sales or marketing team. SparkSense can help draft your acquisition definition from your closed-won deals directly from your CRM. There are no weights to assign and no point values to defend in a spreadsheet, so the definition stays readable by the people who actually know the answer (and can be easily updated as the business shifts).

Write one description per motion

Here's where the setup starts paying off.

A trial account and a renewal account have different litmus tests. "Are these people going to buy" and "are these people going to stay" read from the same source data but mean different things depending on where the account is at in their lifecycle. A drop in weekly active users means something mild during a trial and something urgent sixty days before an enterprise renewal.

Similarly, an SMB account using 100 widgets can read differently than an enterprise account using 100 widgets. Context matters when looking at product data and Trailspark allows you to create as many ICP definitions as you want.

Most teams start with new business, expansion, and renewal, then add lenses over time for things like enterprise vs self-serve, or paid vs free-plan accounts. Each description carries its own rules about which accounts it applies to and its own idea of who belongs in the buying group. An account is evaluated under the lens matching the motion it's in right now, and it moves between lenses as it moves through its lifecycle.

That last part is what makes this genuinely useful rather than merely populated. A rep on a renewal call sees a renewal read, graded against renewal criteria, with renewal roles. A rep on a trial call sees something else entirely from the same underlying activity.

The Your ICPs screen in Trailspark listing three active ideal-customer profiles in priority order. Activation - Active Trial is tagged PLG Acquisition, scores leads, has 4 roles, priority 0. Free Users is tagged PLG Expansion, scores leads, 4 roles, priority 10. Expansion - Current Customers is tagged PLG Expansion, 4 roles, priority 20. A note explains that ICPs higher in the list take precedence when an account matches multiple scope rules. — Three lenses over the same data. Each carries its own motion, its own buying-group roles, and a priority that decides which one wins when an account matches more than one.

A quick way to pressure-test a description

Read it out loud to a rep who works that motion and ask which of their current accounts it describes. If they can name three, you've written a good one.

Step 3: Map the results into your CRM

Next, create the necessary person-level and account-level fields in your Hubspot or Salesforce CRM to map Trailspark's evaluation and reasoning outputs to.

" The reasoning fields are written in natural language so your reps don't have to deconstruct a score to tell the difference between a '78' and an '82' account score. "

The person-level record carries that contact's score, the reason behind it, which lens produced it, and the buying-group role they appear to be playing. The account-level record carries the account's propensity, its own reasoning, which roles are engaged, which are missing, and how complete the committee looks.

A HubSpot contact record for Sandro Hughes, Senior Engineer at Vulcan Ridge Works, open on a custom Trailspark tab. A Contact Scoring card shows a Hot score and a written reason describing strong early-to-mid trial activation with logs and deployment instrumentation, multi-user expansion through team invitations, and feature adoption including MCP and error monitoring, noting activity within the last 14 days and no cancellation signals. A Buying Group card shows stage Forming, engaged role Champion, engagement completeness 20 percent, known completeness 90 percent, and missing role Economic Buyer. A Company Scoring card below shows account propensity high with its own separate reason about APM capacity. — Both levels on one screen. The contact reason names the actual events behind the score, and the company card carries a separate account-level read.

That's the setup. Now let's see how to use it in practice.

Before the call: walk in already briefed

The best prep happens ten minutes before you join, when there's still time to do something with what you learn.

Trailspark allows you to converse with your account data through the MCP server. Simply connect it through ChatGPT or Claude and start your preliminary research:

"Find the account for this [company]. What's changed in the last 60 days? Who are the power users showing the most activity? What is their most used feature?"

The MCP can reference the raw signal history and reasoning behind all the accounts and leads it evaluates. Quick analysis at your rep's fingertips.

A Claude conversation using the Trailspark integration. The reply identifies the account and summarizes what changed in the last sixty days: a trial that converted to a paid subscription, an ICP reclassification from Activation - Active Trial to Expansion - Current Customers, steadily growing usage now over the plan limit driving a high propensity score, and consistent active users with no churn signals. A table lists power users by signal volume with each person's activity pattern, identifying one as the account's technical owner and the others as passive dashboard viewers. A closing read calls the account a healthy expansion opportunity and names the internal champion to engage. — Ten minutes of prep in one answer. The account's lifecycle, who is actually driving usage, and a named person to open with.

Pro tip: Do this before you join rather than during. Waiting on an assistant and trying to read the response while someone is mid-sentence is worse than not having one at all. Use this for quick research and CRM data for reference during the call.

During the call

Your reps keep the CRM record open and reference during the call. The great part about this is that you can customize the view in Salesforce or Hubspot to group the Trailspark data together so the rep doesn't have to hunt for it.

In HubSpot, admins manage this under Record Customization. You can go further than reordering the sidebar and give Trailspark its own tab on the record, sitting alongside About and Activities, with the evaluation and buying-group cards grouped inside it. One click and the rep has everything in one place.

A HubSpot company record for Vulcan Ridge Works, a UK professional services company, with a custom Trailspark tab selected alongside the About, Activities, Revenue, and Intelligence tabs. A Trailspark Evaluation Data card shows account propensity high, an at-risk status of Current, and a written reason stating that APM capacity is 181.8 percent and near the limit, with 10 apps, repeated recent log configuration, and intensive multi-user activity indicating decisive capacity pressure and active product engagement. A Buying Group card below shows stage Forming, engagement completeness 20 percent, known completeness 90 percent, engaged role Champion, and missing role Economic Buyer. — A dedicated tab on the company record. The reasoning is a paragraph a rep can read out loud, not a number they have to decode.

In Salesforce, the compact layout drives the Highlights Panel at the top of the record page. You get up to ten fields, with roughly the first seven rendering in Lightning depending on screen width. Build a custom compact layout, since the system default can't be edited.

Either way, put propensity and the written reasoning at the very top. Everything else on the record can stay where it is.

What this changes about the first ninety seconds

Instead of a warm-up question, the rep opens with something specific: "I saw your team got SSO configured last week. How did that go?" The conversation starts three minutes ahead of where it usually does, and the prospect feels like they're getting the proper attention.

You can also use the information from the research phase to ask about missing buying group roles that Trailspark surfaced.

After the call: close the loop while it's warm

If you're using Zoom or Granola to transcribe your calls and take notes (and why wouldn't you be by now?), the call just produced a new artifact that you can immediately put into action.

Trailspark can tell you an account went quiet in the product three weeks ago. The transcript can tell you the champion mentioned a reorg on the last call. Neither is the full story alone, and the combination explains the account in a way that no single source does. Use that data to be proactive and determine who you might need to bring into the conversation from that reorg.

You could take it a step further and point an agent at both the Trailspark data and the meeting transcription to help facilitate this. Trailspark's MCP endpoint for the evaluation and the buying group, your CRM for deal state and history, and your transcript tool for what was said. Ask it to produce a short status per account, refreshed weekly.

For sales leadership, that's a pipeline review that writes itself, grounded in evidence rather than rep optimism. For the rep, it's a standing brief on every account they own.

Tune what gets delivered

A few settings control how much shows up, and knowing them lets you dial the experience to what your team wants.

Score tiers decide which accounts get pushed. The default is deliberately conservative and sends your hottest accounts only. Widen it when you want warm accounts reaching reps too, which most teams do once they've seen this working.

Buying group push adds the role and coverage fields. Turn it on and each contact carries the role they appear to be playing, which is what lets a rep tell the champion from the security reviewer before anyone speaks.

Create if not found decides what happens when a meeting attendee has no CRM record yet. Turning it on is especially valuable in a product-led motion, where a good share of the buying group arrived through the product and never filled out a form. This is the setting that gets those people onto the record in the first place.

Every destination keeps a push log per record, so you can always see what was written and when.

Quick-Start Checklist

Connect one signal source - Product events or your CRM. Confirm signals are landing before tuning anything else, since everything downstream reads from them
Write your first description - Pick one motion, describe a good account in plain sentences, and have a rep who sells that motion read it back to you
Add a second lens - Renewal or expansion is the usual next one, and the moment the two disagree about an account is the moment this gets interesting
Create your CRM fields - Person-level and account-level, with a long-text field for the reasoning at each level
Map both levels - Score, reason, scoring lens, and role on the person. Propensity, reason, roles engaged, and roles missing on the account
Set your delivery preferences - Score tiers, buying group push, and create-if-not-found
Preview, then enable - See exactly what would be written before it is
Move the reasoning to the top of the record - HubSpot record customization, or a custom Salesforce compact layout
Hand one rep a read-only agent credential - Have them use it before a call and again after, then tell you which one they'd keep

Start with one motion and one rep. It sells itself from there, and the second lens takes a fraction of the time the first one did.

What to Read Next

For the thinking behind evaluating the account rather than the individual, Account Scoring vs Contact Scoring covers why the account is the right unit for B2B and how contact-level signals feed it.

The 2026 Guide to AI Lead Scoring walks through the wider framework, from defining your outcome to choosing signals and evaluating tools.

And Why Rules-Based Lead Scoring Breaks Down goes deeper on why a written reason is the thing that gets a score used.

Trailspark evaluates product usage, demand gen signals, and ICP fit together in a single full-context assessment, explains every score in plain language, and writes the result back to the records your team already works in. Sign up free →

Upleveling Your Sales Calls: How to Use Better Account Context Before, During, and After the Call