Sable c eated Aida , a AI employee, who helps busi esses show all buye s what a p oduct ca do fo them, i stead of just telli g them. Aida leads live p oduct expe ie ces, a swe s complex questio s, a d guides buye s th ough p oducts i eal time while measu i g e gageme t, gathe i g feedback, a d su faci g i sights f om eve y i te actio . The b eakth ough e abli g this is what Sable calls I te active I tellige ce, the combi atio of eal-time b owse avigatio , visio , voice, a d video that allows AI to see, click, explai , a d collabo ate with use s i a sha ed e vi o me t. U like a chatbot waiti g to be p ompted, Aida pe ceives what’s o sc ee , speaks atu ally, a d avigates the p oduct di ectly, so custome s ca expe ie ce the same dy amic, atte tive sessio delive ed by a huma employee, at a y hou , i a y la guage.
This was ot possible eve a few mo ths ago. Imp oveme ts i visio a d compute -use models, combi ed with adva ceme ts made by Sable’s team, ow e able Aida to pe fo m live b owse actio s o dema d a d co ti uously pe ceive you sc ee a d espo d i eal time. Aida also has its ow wo kspace, called a LiveBox, desig ed fo live, simulta eous custome i te actio . A LiveBox is a vi tual machi e that allows Aida to demo st ate a p oduct’s full capabilities to buye s, helpi g them u de sta d what it ca do fo them.
“By combi i g compute use, visio , a d voice, we’ve c eated a full AI employee with a deg ee of auto omy these systems have eve had. While lo g-ho izo age ts complete backg ou d tasks, Aida i te faces di ectly with buye s o you behalf, e d-to-e d, without a huma i the loop,” said Nim Ravid, CEO of Sable. “Achievi g the low-late cy a d high-accu acy i te actio equi ed fo a atu al co ve satio with a AI employee mea t ove comi g a set of f o tie AI challe ges.”
While chatbots a e effective fo tasks like d afti g emails o summa izi g docume ts, they a e ot the ideal i te face fo evaluati g o pu chasi g p oducts o li e. They push visual co te t to the backg ou d, co st ai i te actio s to tu -based p ompts i stead of di ect ma ipulatio , a d do ‘t allow fo simulta eous collabo atio . The combi atio of Sable’s I te actio I tellige ce a d LiveBox i t oduces a ew i te actio pa adigm that e ables busi esses to lau ch a AI employee who is i fi itely k owledgeable, patie t, a d always p ese t, delive i g pe so alized, white-glove buyi g expe ie ces that we e p eviously ese ved fo o ly a small f actio of custome s.
Today’s buyi g expe ie ce is quite the opposite, leavi g custome s f ust ated. Pu chasi g AI a d softwa e p oducts usually mea s custome s schedule a qualificatio call with o e pe so , sit th ough a sepa ate demo with a othe , meet agai with a accou t executive to close, a d get ha ded off o ce mo e fo o boa di g, losi g co text with eve y ew i te actio , havi g to epeat themselves at eve y step, a d st uggli g to get immediate a swe s. Sable eplaces that disjoi ted expe ie ce with
Deployi g Aida is simple a d equi es o i teg atio . Sable e ables compa ies to teach Aida thei best p actices a d i stitutio al k owledge fo positio i g p oducts a d leadi g custome co ve satio s. Sales call eco di gs, i te views with top pe fo me s, docume tatio , a d ma keti g mate ials a e t a sfo med i to a u ified e te p ise custome co text g aph, called the B ai . F om the e, Aida co ti uously lea s by ide tifyi g k owledge gaps, captu i g successful i te actio patte s, a d efi i g both the accu acy of its espo ses a d the quality of its co ve satio s. Withi weeks of deployme t, the B ai maps a o ga izatio ‘s u ique expe tise i to a st uctu ed, co ti uously evolvi g k owledge lib a y.
“The defi i g comme cial oppo tu ity i AI comes dow to how fast compa ies ca adopt what the f o tie has al eady built,” said Shau Magui e, Pa t e at Sequoia. “Sable has c eated a di ect path f om f o tie capability to how busi esses actually each thei custome s: a vi tual employee that ca see, click, avigate, a d explai i eal time.”
D awi g o decades of expe ie ce buildi g e te p ise softwa e, Pala ti co-fou de Joe Lo sdale said, “Fo decades, compa ies have elied o sales teams, solutio e gi ee s, a d fo wa d-deployed e gi ee s to help custome s u de sta d what p oducts ca actually do fo them. B eakth oughs i eal-time compute use a d visio ow allow fo the automatio of la ge pa ts of custome -faci g wo k. Sable AI’s employee is e abli g compa ies to scale i ways that we e totally impossible eve a yea ago.”
Sable was fou ded by Nim Ravid, Leo Che , Li da He, a d Itama Rocha, who met at Ha va d U ive sity a d co ducted AI esea ch i post-t ai i g, ei fo ceme t lea i g, a d multimodality. P io to fou di g Sable, membe s of the team wo ked at SpaceX, Google, Meta, a d Togethe AI. The b oade team today i cludes AI esea che s, fo me qua titative t ade s, a d I te atio al Math Olympiad wi e s.
