A manual RFP bank that capped how many deals the team could chase
Before SiftHub, Capillary's presales team ran every response out of a manual RFP bank. A person searched for answers by hand, then drafted, rewrote, and formatted them one by one.
"Someone used to go search for answers manually, fill those answers in, and edit those answers in the right English," said Sanchit Singla, Senior Director and Head of Presales for Asia at Capillary Technologies. "A lot of manual work around drafting and documenting used to happen, which used to also restrict us in terms of the number of RFPs we can participate in a quarter."
Quality was the problem that personally cost Sanchit the most time. "Even if the best of my presales guys can write the answers, they will make spelling mistakes, they may not write the right grammar, or they may not find the right answers," he said. "Someone may write a paragraph, someone may write bullet points, and someone may not remember the right example or may not have access to the right number that is available in the repository." Getting those answers in the first place meant chasing security and product teams across email and Jira, coordination he described as "very difficult for someone to manage."
Then there was the deadline problem. "I don't have visibility into how many of these have been completed, not completed, 60% done, 40% done," Sanchit said. "Only two days before the deadline is when I get to hear the bad news or the good news."
One platform that stitches every source into a single answer
Capillary had been looking for a fix for two years. The team ran internal presentations on why it needed the software, evaluated a long list of vendors, and had already run proofs of concept with two competitors before SiftHub was introduced. The first round judged on fill rate alone. The second went deeper. "We found that the other tools that we were looking at were actually probably hallucinating for more than 60% of answers," Sanchit said. "So while it was a 9 on 10 fill rate, the answers were hardly contextual." SiftHub joined the same proof of concept, and the result decided it. "We were very impressed by the fact that the answers were very contextual, and if it did not give me an answer, it was only because it did not have that context."
Today every RFP starts with a 15 to 20 minute planning call to flag red and amber issues across product, compliance, legal, and commercial, and to decide whether to bid. The team then dissects the RFP, runs SiftHub Autofill for the first pass, and uses the AI Teammate for the questions Autofill leaves open and for on-the-fly refinements. What holds the answers together, in Sanchit's view, is orchestration. "The way SiftHub orchestrates my product documents, my past RFPs, my GTM body of knowledge, all together and stitches it together for me is what is even more helpful," he said. "As a person, I may only be able to check my RFP bank, or I may just check some other resource. What it beautifully does is it orchestrates the answer across all these resources and puts it into one single answer for me." Sources that used to sit in separate systems and stray Excel files now live in one place: "All of that has been unified on one system for me."
The collaboration problem solved itself in a way Sanchit did not expect. His information security team now works in SiftHub directly. "I told her, okay, can we upload them to SiftHub? Just give it to us, we will upload it," he recalled of a recent exchange over a sheet of security answers. "She said, no, I know how to upload it on SiftHub. She went, and she did it herself." His read: "That's something that really made it more empowering for the entire team." Capillary also worked with the SiftHub team to build Slack and email notifications that flag pending projects, so final versions get saved back into the platform and leadership can see project status without asking.
Work that took a strong performer five days now takes a good performer one
"The person used to work on the deck, the Excel filling up, and drafting the answers. It used to take at least four to five days, and I'm talking about a smart fellow," Sanchit said. "For someone who's challenging himself to step up into the role, it would be like a week's task with a lot of night hours spent and a lot of bad feedback there. That has now come down to less than a day for a good performer, and at most two days." Review used to add a day or two on top. "For review, it doesn't take more than a day now, because at least the answers are in the right shape."
Sanchit's own review load fell with it, from "roughly in a week 8 or 10 hours reviewing RFPs" to about three. Across the team, turnaround time is down 80%, numbers Capillary has published internally in its presales updates and QBRs. Answers requiring zero edits have climbed from a realistic 50 to 60% a year ago to 80 to 85% today.
The freed capacity went straight back into the pipeline. Capillary now runs 2x the number of RFPs with the same headcount. "In the last year, we had almost a rejig of our internal team, and almost everyone in the team today that you talk to is less than a year old into the system," Sanchit said. "You can imagine, for someone new, understanding the platform was obviously going to be difficult. But they were still able to deliver their RFP answers with the help of SiftHub, because they had something to at least do a starter for themselves." For a leader who has watched internal projects get cut for failing to prove themselves, that was the test: "Such projects generally in our kind of an organization get killed out in the first year itself. Because if something is not showing value in the first year, it's best to kill it. SiftHub has already crossed that mark for us."
Would Capillary recommend SiftHub?
"It is a highly competent technology platform. It is one of the only platforms that understands your business and is able to stitch it well together with you as an AI product. The values in the company are very strong, very customer-first, very outcome-focused, which is a quality that is there in very few companies. Regular updates, regular follow-ups, asking for more. Very few companies have KPIs on those."
.svg)
.avif)
