Kiba grew ARR by 133% in eight months with a fully automated outbound system

Kiba was doing everything manually: LinkedIn prospecting, email research, personalisation, sending, follow-ups, CRM updates. Their two-person sales team could reach 25 prospects per day and spent less than 20% of their time in actual sales conversations. Revenue growth had flatlined at under 5% quarter-on-quarter.

The workflow looked roughly the same every morning. One rep would open LinkedIn Sales Navigator and spend two hours scrolling through profiles, copying contact details into a spreadsheet, and cross-referencing company information on Crunchbase. The other rep would draft personalised emails in Gmail, manually pulling in details about each prospect's company, recent funding rounds, job postings, product launches, to make the outreach feel relevant. By the time both reps had their lists ready and emails queued, it was nearly midday. The actual sending, tracking, and follow-up sequences were managed through a patchwork of browser tabs, a shared Google Sheet, and HubSpot entries that were frequently out of date.

Follow-ups were the biggest casualty of this manual process. Prospects who didn't respond to the first email rarely received a second touch within the optimal window. The team had set reminders in their calendars, but these were inconsistent and easy to miss during busy weeks. When a prospect did reply, the rep often had to re-read the original email thread and cross-check the CRM to remember the context. CRM hygiene was poor, deal stages were updated sporadically, notes were sparse, and pipeline reporting was unreliable.

The result was a ceiling that Kiba could not break through. Despite having a strong product-market fit and healthy close rates when they got on calls, the top of the funnel was starved. Leadership had considered hiring a third rep, but unit economics did not support it at their current deal size. They needed a way to multiply output without multiplying headcount.

We built a fully automated outbound pipeline that replaced every manual step from prospecting to CRM update. The project kicked off with a two-week discovery phase where we mapped Kiba's existing workflow in detail, documented their ideal customer profile across twelve firmographic and behavioural attributes, and audited their CRM data to establish a clean baseline.

The first module was an AI sourcing agent that runs on a nightly schedule. It queries LinkedIn Sales Navigator, Crunchbase, and three industry-specific directories using Kiba's ICP criteria, company size, funding stage, technology stack, hiring patterns, and geographic focus. The agent pulls company profiles and contact data, deduplicates against existing CRM records, and stores enriched profiles in a staging table. Each night's run typically surfaces 300 to 400 net-new contacts.

The second module is a lead scoring model built on a weighted combination of fit signals and intent signals. Fit scoring uses firmographic data, revenue band, employee count, industry vertical, and tech stack overlap. Intent scoring layers in real-time signals: recent job postings for roles that suggest a need for Kiba's product, new funding announcements, leadership changes, and website technographic shifts detected via BuiltWith data. Leads are scored on a 0-to-100 scale, and only those above a threshold of 65 enter the sequencing pipeline.

The third module handles personalised outreach. GPT-4 generates three-email sequences for each qualified lead, using structured prompts that incorporate the company's recent news, funding status, job postings, and any mutual connections. Each email is reviewed by a tone-and-compliance filter before entering the send queue. Sequences execute automatically through a connected Instantly account, with send times optimised based on the prospect's timezone and historical open-rate data.

Finally, a bi-directional CRM sync ensures that every event, email sent, opened, replied, bounced, meeting booked, writes back to HubSpot in real time. Deal stages advance automatically based on engagement milestones. The sales team opens their CRM each morning to a prioritised list of warm leads with full context, ready for a conversation.

Outreach scaled from 25 to 250 qualified prospects per day within the first three weeks of deployment. The sourcing agent consistently surfaced 300-plus net-new contacts nightly, and the scoring model filtered these down to the highest-fit, highest-intent subset. The sales team no longer spent their mornings on research and list-building, that entire block of work simply disappeared from their calendars.

Sales team time on actual calls and conversations increased from 20% to 75%. Both reps reported that the quality of their conversations improved as well, because the AI-generated email sequences had already established context and relevance before the first call. Prospects arrived on calls having engaged with personalised content about their specific pain points, which shortened the discovery phase and moved deals through the pipeline faster.

Kiba grew ARR by 133% over eight months. The growth was not evenly distributed, the first two months were spent refining the ICP scoring model and tuning email templates, with modest gains. Months three through eight saw compounding returns as the system learned from engagement data and the scoring model became more precise. Two new enterprise accounts, each worth more than five times Kiba's average deal size, were sourced entirely by the automated pipeline, neither rep had any prior relationship with those companies.

CRM hygiene improved dramatically as a side effect. Because every touchpoint was logged automatically, pipeline reporting became accurate for the first time. Leadership could see exactly how many prospects were in each stage, what the average time-to-close looked like, and where deals were stalling. This visibility enabled better forecasting and more informed decisions about pricing and packaging. The system continues to run autonomously, and Kiba has since expanded their ICP to target a second vertical using the same infrastructure.