#116 Sales Forecasting in the Age of AI – with Janis Zech & Philipp Stelzer (WeFlow)

In this host-only episode of the RevOps Lab, Janis and Philipp take stock of what three years of building a forecasting tool — and hundreds of conversations with sales leaders, RevOps teams, and CROs — have taught them about getting to a reliable, repeatable forecast number. They unpack why forecasting is a process (not a number), why AI only works on top of a clean data foundation, and how the best companies combine roll-up, dynamically weighted, and AI-predicted forecasts into a single operating cadence.

We cover:

• Why forecasting accuracy is the output of a well-run sales org, not the input

• The operating cadence: weekly meetings, deal reviews, and stakeholder alignment that make forecasting work

• Building the data foundation: activity capture, multi-threading signals, conversation intelligence, and CRM autofill

• Why deal hygiene and shared qualification criteria (SPICED, MEDDIC) are non-negotiable before you forecast

• Splitting the forecast into new logo, expansion, and renewal — and why bookings ≠ consumption

• The three-pillar forecast: dynamically weighted + bottom-up roll-up + AI prediction (with corridors, not single numbers)

• How to run a roll-up motion: baseline vs. best case, rep forecast vs. independent manager forecast

• Why running roll-ups in spreadsheets breaks down at 50+ reps

• Calculating dynamic stage probabilities by rep tenure or team — and when it's worth doing

• Why AI predictions are only as good as the data foundation underneath them

Links:

• Janis Zech on LinkedIn: https://www.linkedin.com/in/janiszech/

• Philipp Stelzer on LinkedIn: https://www.linkedin.com/in/philippstelzer/

• WeFlow: https://www.getweflow.com

• WeFlow RevOps resources: https://www.getweflow.com/revops

• Join the RevOps Chat Community: https://www.getweflow.com/community

• Subscribe to the RevOps Letter: https://www.getweflow.com/revops-letter

• Operating Cadence master deck: ping Janis or Philipp on LinkedIn to request

Chapters:

• (00:00) Intro: Has AI fundamentally changed forecasting?

• (01:34) Why forecasting is a process, not a number

• (02:25) What an operating cadence actually looks like week-to-week

• (04:23) The data foundation: why CRM alone isn't the system of truth

• (06:53) Anchoring deal conversations on a shared qualification methodology

• (08:42) How AI adds an objective layer through automated capture and CRM autofill

• (09:29) Stage entry/exit criteria and deal signals for deal health

• (13:06) Why "comparable deals" matters as you scale past 50 reps

• (14:23) Forecasting as the end result of a well-functioning sales org

• (16:05) Splitting forecasts: new logo, expansion, renewal, bookings vs. consumption

• (17:22) The three-pillar forecast: dynamically weighted, roll-up, AI prediction

• (18:50) Roll-up forecasting: baseline vs. best case, rep vs. manager numbers

• (24:07) Anatomy of a roll-up: hierarchy, gap-to-quota, pipeline coverage, deal-by-deal

• (26:43) Why spreadsheets break down for roll-up forecasting at scale

• (28:07) AI prediction models: aggregate vs. deal-by-deal scoring

• (29:02) Dynamically weighted forecasts and rep-level stage probabilities

• (32:44) Why AI predictions only work on top of clean data foundations

• (34:27) Book recommendation & close