Every sales leader has lived the same quarter-end ritual: a forecast number that looks confident on Monday and evaporates by Friday. The problem was never the number itself. It was that nobody could tell you why the number was what it was, or whether it reflected what customers actually said, or just what reps hoped.
We built Airspeed Forecasting to close that gap. Here’s what makes it different, and why we think it sets a new bar for the category.
A number is not an answer
Gong used to give us a close probability, but it was just a number. Airspeed gives us a close probability and it’ll actually give you a breakdown of why that percentage is the way it is.
Dakota Kummer, Sr. Sales Manager at Peek
That distinction is the whole point. Most forecasting tools hand managers a percentage and leave them to guess at the story behind it. Ours shows the reasoning, the trend, and the risk, so the score becomes something a rep can act on, not just a verdict they have to defend.
Built on what was said, not what was typed
Most forecasting tools lean on CRM fields that reps fill in by hand, and those fields are optimistic at best and stale at worst. Our close-probability scores are derived from real call and meeting content: what was actually said, not what someone remembered to log. The forecast reflects ground truth instead of wishful thinking.
On top of that, the AI infers deal-qualification signals (MEDDIC, BANT, SPICED) straight from the conversation: budget signals, decision-maker identification, timeline and urgency, and the real pain points driving the deal. Instead of relying on reps to self-report, the model reads the room. And because every new call re-scores the deal and syncs with HubSpot or Salesforce automatically, the forecast stays current without anyone touching a spreadsheet.
Three things a dashboard number can’t give you
Reasoning you can drill into. Every open deal gets a probability percentage, the reasoning behind it, and a trend: improving, declining, or flat. Managers and reps see exactly what’s holding a deal back.
Loss intelligence that compounds. For closed-lost deals, we automatically surface loss-reason categories and plain-language summaries, grounded in the calls. Over time, patterns emerge and feed directly back into coaching and process.
Risk, before it’s too late. Stalled activity, missing stakeholders, unaddressed objections: these get flagged proactively. You get early warning while there’s still time to act, not a post-mortem at quarter close.
How to forecast truthfully
Getting this right was harder than it looks, and the hard parts taught us what honest forecasting actually requires. A few principles we now build around:
Don’t punish accuracy. Most close dates are arbitrary, set by internal pressure rather than customer signals. Tools that penalize a rep for moving a close date end up punishing the very honesty you want. A truthful forecast has to make room for the real “most likely” date.
Forecast the period, not eternity. “Will this deal ever close?” is the wrong question. Leaders are held accountable to the quarter, so probability has to be period-aware: the odds it closes this period, which is the only version that maps to how CROs are measured.
Judgment beats box-checking. As forecasting expert Tom Andrews put it, hitting 80% of MEDDPICC and then chasing the rest often has no material impact on the outcome. Human judgment about forecast category matters more than framework completeness, and the AI weighs the conversation accordingly.
Never be the rep’s enemy. A forecasting AI cannot be a surveillance tool. Our rule, drawn straight from CRO advisors: never show a manager anything the rep can’t also see. The reasoning is surfaced to both, equally.
Does it work? In evaluations across real customer organizations, the model’s predictions proved more accurate than reps’ own forecasts on their own deals. That’s the bar we hold ourselves to: an AI layer only earns its place by beating human judgment on real outcomes. Co-founder and CTO Devang Agrawal walked through the approach live on stage; read the recap in Building Agents With a Brain.
Live webinar: Your Sales Forecast Is Lying to You
Join us on July 28 with Doug May, SVP at Harness, and Thang Nguyen, VP of Sales at Airspeed, for a candid session on building forecasts your team believes in, and why the honest number is the accurate one.
The takeaway
Forecasting isn’t a static snapshot. It’s a living, conversation-backed view of your pipeline that gets sharper the more your team uses it, closing the gap between what reps say is happening and what’s actually happening in the deal.
