Backstory retiered 141 accounts in 3 days using AI signals, not spreadsheets
Haya Kamola, who leads customer success at Backstory, walked through an account tiering project at SaaStr AI Day that most GTM leaders will recognize: take the customer base, figure out which accounts deserve the team's time, redistribute accordingly.
The board set a short deadline. 141 accounts. Define best customers, measure everyone against that, produce a framework execs can act on.
Previous version: her team plus four others, working a full quarter. This version took three to four days.
The definition came before the data
Kamola started by asking account teams and leadership to describe what made one or two specific customers different. Not longest-tenured, not largest.
What landed: customers who treated Backstory as core infrastructure, built systems around it, planned five years out with the product at center, cared about roadmap, kept finding new use cases.
Everything else got measured against that. Without it you score accounts against whatever fields happen to be populated in your CRM.
Most signals did not exist yet
Stripping down the golden customers produced two sets of characteristics.
Internal to customer: GTM process maturity, tech stack mix, AI maturity and velocity, partner motions.
Specific to relationship: deployment velocity (landed small, expanded fast), executive visibility (data used by execs for decisions), total addressable market within account, remaining white space.
A lot of that sat in silos. Several signals had never been measured repeatably across the full base, so gaps got filled before analysis started.
AI maturity was the example. Backstory had an internal five-level framework covering culture, investment, tech stack, talent, willingness to engage on hard problems. Maintaining it meant asking account teams to categorize customers by hand.
They built it as a signal that runs systematically: a prompt executed against every account, pulling specific CRM fields, public company information (AI product launches, new investments), and full conversation history (emails, meetings, Slack). Output: maturity level per account plus reasoning.
Same for tech stack mix, where pre-sales scorecards from two to three years ago were stale and current picture sat in conversations nobody had systematically read.
Cross-functional data pull became four connectors
This part used to consume the quarter. A project owner going to product, BI, finance, other functions to collect TAM, health, revenue, renewal rates, delivery risks, feature requests, adoption levels for every account.
Kamola asked nobody. Four connectors covered it:
- Amplitude for utilization and usage data
- Atlassian and Jira for years of logged customer feature requests and gaps
- Backstory's own MCP for full conversation and engagement history
- Slack, because Backstory runs an internal channel for every single customer, and those channels are where account strategy, next steps, risks get called out first
The Slack input is the one most teams could copy tomorrow. An account team's internal dialogue is usually the earliest, most candid read on a customer, and it almost never makes it into any structured system.
Only manual step: exporting the customer list from Salesforce. Account name, executive engagement level, predicted health, AI maturity signal, upcoming renewal date, renewal ACV.
Four iterations to narrow eight signals to four scoring buckets
The analysis runs as a defined sequence rather than a single prompt. Ingest and normalize the CSV so accounts match across sources. Start with conversation history as primary source of truth for recent engagement and risks. Layer in usage data, feature requests, tech stack signals. Apply the AI maturity score. Calculate white space.
One signal was scoring backwards in the first iteration, which they caught because the output ranked accounts they knew were tier one as tier three. Adjusted weights, ran again.
Four rounds total to get the tiering framework the exec team could act on.
What this means for account planning
Account tiering, customer segmentation, prioritization frameworks: most sales and CS orgs run some version of this exercise annually or when leadership changes. Traditional approach is cross-functional, manual, slow.
Backstory's version automated the data collection (connectors replace meetings), built repeatable signals for characteristics that previously required judgment calls (AI maturity, tech stack analysis), and used structured conversation history (Slack channels, email, meetings) as a primary input rather than an afterthought.
The workflow is portable. Four connectors (usage tool, project management, conversation platform, internal comms) plus one CRM export is table stakes for most B2B shops. The difference is treating internal account team dialogue as data and building signals that run systematically instead of asking humans to score accounts by feel.
Worth noting: Backstory is a revenue intelligence company, so this is also a product case study. But the account tiering logic and connector strategy work regardless of tooling. The underlying shift is moving from periodic cross-functional projects to always-on signals that update as customer behavior changes.
Salesforce Sales Cloud has native account scoring and tiering features, but most orgs still run spreadsheet exercises because the signals that matter (conversation sentiment, tech stack fit, AI readiness) do not auto-populate in standard fields. Backstory's approach bridges that gap by making non-standard signals systematically available.
If your account planning process still takes a quarter and requires five teams, the bottleneck is probably data assembly, not analysis. Fix the connectors, define the signals, write down the workflow. Three days is achievable.