AI boosts forecasting

Most sales forecasts still break in the same two places: deals that look real but never close, and deals that close later than the quarter says they will. Gartner says only 7% of sales organizations reach forecast accuracy above 90%, and 69% of sales operations leaders say forecasting is getting harder. (gartner.com) (demandgenreport.com) The pitch behind the new wave of sales artificial intelligence is simple: stop asking managers to guess, and start reading the trail each deal leaves behind. Gartner’s September 25, 2024 research says artificial intelligence can improve data capture, predictions, and insights while reducing seller burden. (gartner.com) One piece of that is predictive lead scoring, which works like a credit score for a potential customer. Microsoft says its Dynamics 365 Sales model assigns each lead a score from 0 to 100 based on signals from the lead, contact, and account, so teams can rank which names are more likely to convert. (learn.microsoft.com) That changes the forecast because a pipeline with 200 leads is not really 200 equal chances. Microsoft’s documentation says the model surfaces the top factors behind each score, which lets a manager separate a likely buyer from a name that just filled out a form. (learn.microsoft.com) The second piece is next best action, which is software that tells a seller the most useful next move on a live deal. Microsoft’s 2026 release plans describe next best actions that highlight which leads need attention and recommend steps like follow-up, qualification, or unblocking the agent inside the seller workflow. (learn.microsoft.com 1) (learn.microsoft.com 2) That matters most in long business-to-business sales cycles, especially in hardware, where one order can take 6 to 12 months to move from first meeting to signature. In those cycles, a missed technical review in May can turn into a slipped quarter in September, even if the deal still looks “green” in the customer relationship management system. (salesforce.com) (hubspot.com) That is why revenue leaders keep coming back to pipeline velocity, a metric that turns a messy sales funnel into one number. Salesforce and HubSpot both define sales velocity with the same four inputs: number of opportunities, average deal value, win rate, and sales cycle length. (salesforce.com) (hubspot.com) The formula is straightforward: opportunities multiplied by deal value multiplied by win rate, then divided by the length of the sales cycle. Salesforce says the result tells you the revenue you can expect to generate through that cycle, which is why a slower cycle can wreck a quarter even when pipeline dollars look healthy. (salesforce.com) For a hardware company, that denominator matters more than in fast software sales because every extra month drags down the whole machine. A team with 40 deals, a $150,000 average deal value, a 25% win rate, and a 300-day cycle produces far less near-term revenue than the same team with a 210-day cycle, even if nothing else changes. (salesforce.com) (hubspot.com) This is where artificial intelligence stops being a chatbot story and becomes an operations story. If predictive scoring raises the quality of opportunities entering the funnel, and next best action cuts idle time inside each deal, both changes feed directly into the same velocity equation by lifting win rate or shortening cycle length. (learn.microsoft.com 1) (learn.microsoft.com 2) (salesforce.com) The claim that forecast accuracy can jump by 30% to 50% is best read as a practical operator’s estimate, not a universal benchmark published by one neutral standard. What the broader market does show is a large accuracy gap, a growing set of tools built around scoring and recommended actions, and a clear math path for why those tools would help most in slow, high-value sales motions. (gartner.com) (learn.microsoft.com 1) (learn.microsoft.com 2)