Most account tiering projects take too long. That’s not an opinion – it’s a resource problem that quietly stalls pipeline building for months at a stretch.
At SaaStr AI Day, Haya Kamola, who leads customer success at Backstory, described how her team retiered their entire customer base in three days. The same exercise had previously required five teams and a full quarter to complete. The difference wasn’t headcount or budget. It was a combination of AI connectors, custom signals, and a willingness to iterate quickly rather than waiting for a perfect data set before moving. That’s a meaningful shift in how RevOps and CS teams can think about segmentation work going forward.
Why Account Tiering Keeps Getting Deprioritised
Tiering is foundational. It determines where your team spends time, how you structure coverage, and ultimately what your Net Revenue Retention (NRR) looks like at the end of the year. And yet it almost always gets pushed back.
The reason is straightforward. Traditional tiering analysis requires pulling data from multiple sources, aligning on scoring criteria across sales and CS, normalising signals that don’t map cleanly to each other, and then getting leadership sign-off on a framework that will be outdated within two quarters anyway. It’s a slow process built for a slower world. Teams don’t abandon tiering because they think it’s unimportant – they deprioritise it because the operational cost feels too high relative to other fires.
This is exactly where AI-assisted analysis changes the equation. Custom signals – things like product usage depth, support ticket frequency, and expansion history – can now be weighted and re-weighted in hours rather than weeks. Iteration becomes cheap, and that changes the psychology of the whole exercise.
What Good Tiering Actually Feeds Into
Tiering isn’t an end in itself. Done well, it feeds four things your GTM motion depends on.
- Coverage decisions: Which accounts warrant a dedicated CSM versus a scaled or pooled model.
- Expansion prioritisation: Where your team should focus upsell and cross-sell energy based on fit and growth potential.
- Churn risk flagging: Tier 1 accounts with declining engagement signals need different intervention than Tier 3 accounts with the same pattern. Treating them identically is expensive – see our Churn Rate glossary entry for context on how this affects overall revenue health.
- Pipeline sequencing: For new business teams, understanding which segments of your Ideal Customer Profile (ICP) are converting fastest helps you weight your sales pipeline more accurately.
The Backstory example matters here because it demonstrates that speed doesn’t require sacrificing rigour. Four rounds of iteration in three days is more analytical work than most teams do across an entire quarter-long tiering project. AI compresses the feedback loop – not the thinking.
The GTM Stack Is Getting Rebuilt Around Speed
Backstory’s experience isn’t happening in isolation. The broader market is moving in the same direction.
Five9 raised its full-year AI growth outlook to 60% following a strong Q2, and closed its largest Google Marketplace deal to date. That kind of signal from a contact center vendor matters to GTM teams because it reflects genuine enterprise appetite for AI-assisted customer interactions at scale – not just experimentation. When contact center AI adoption accelerates, the data those systems generate about customer sentiment, escalation patterns, and resolution rates becomes a legitimate input for CS tiering models.
BrandJet AI’s acquisition of lead generation platform IGLeads points to the same consolidation pressure. Teams that previously managed separate tools for prospecting, scoring, and outreach are being pushed toward more integrated stacks. That’s relevant for anyone building a go-to-market motion right now – the platforms are converging, and your process design should account for fewer handoffs between systems, not more.
Five9 raised its full-year AI growth outlook to 60% after beating guidance on both revenue and EPS in Q2 2026, and closed its largest-ever Google Marketplace deal. – CMSWire, August 2026
What’s harder to see, but equally important, is the traceability trend showing up in industrial sectors. The label printer applicator market is projected to reach $1.60 billion by 2033, growing at 4.5% CAGR, driven largely by compliance requirements and the need to reduce mislabeling errors through inline automation. It’s a physical-world analogue to what’s happening in GTM data: the cost of incorrect classification is rising, and automation is the practical response. Mislabeled accounts in your CRM – wrong tier, wrong owner, wrong stage – carry real revenue consequences that compound over time.
Building a Faster Tiering Process Your Team Will Actually Use
The operational lesson from Backstory isn’t “buy an AI tool and run it once.” It’s about changing how your team treats tiering as a recurring motion rather than an annual project.
Start with three to five signals that genuinely predict expansion or contraction in your customer base. Don’t try to boil the ocean on day one. Product engagement data, contract size relative to company headcount, and support-to-value ratio are usually strong starting points. Layer in firmographic signals from your ICP definition to add context around growth potential.
Then build for iteration. Your first tier output will be wrong in places – that’s fine. The goal of round one is to surface the accounts that clearly don’t belong in their current tier, the ones everyone already suspects are misclassified but nobody has had the bandwidth to formally move. Fix those first. Build confidence in the process, then run a second pass with refined weights.
If you’re evaluating tools to support this kind of workflow, the CRM Tools Directory is a good place to compare options across data enrichment, CS platforms, and RevOps tooling. For teams new to formalising their segmentation approach, the CRM Guides section covers account scoring frameworks in practical detail.
The teams winning on Annual Recurring Revenue (ARR) growth right now aren’t doing more tiering analysis. They’re doing faster tiering analysis and acting on the output before the data gets stale.
Speed is the strategy. Everything else follows from it.