Outcome-Based AI ProjectSales / B2B·May 2026

A Centralised AI Hub for Field Sales: Follow-Ups, LinkedIn Content, and Objection Handling — All from Your Phone

A B2B sales team operating in industrial sectors was losing context between client visits, sending generic follow-ups too late, and leaving valuable field insight unused. Kovil AI built a mobile-first AI portal with three integrated capabilities: a per-client AI assistant for personalised follow-ups, a voice-to-LinkedIn content engine, and an AI-enhanced objection response database.

Faster follow-up emails sent

From days to within the hour

60s

Voice memo to LinkedIn post

Transcribed & drafted automatically

100%

Client context preserved

Per-workspace relationship memory

Mobile-first

Voice across all tools

No laptop required

Client type: B2B Field Sales Team (Industrial Sector)
Timeline: 6 weeks
Team: 2 engineers

Tech Stack

Next.js 15TypeScriptSupabaseGPT-4o / ClaudeOpenAI WhisperLinkedIn APIHubSpot Webhooks

"I used to send follow-ups two days after a meeting if I remembered. Now I'm sending personalised emails from the car park before I've driven away. The LinkedIn content alone has generated three inbound conversations this month."

Senior Account Executive, Industrial B2B Sales

The Brief

The client sells into industrial and manufacturing sectors — a relationship-driven environment where field reps spend most of their time on customer sites, not at desks. They came to Kovil AI with a specific problem: their reps were losing context between client visits, generic follow-up emails were damaging relationships they had spent months building, and valuable insight gathered in the field was never making it back into the business in a usable form.

They wanted a single tool a field rep could use from their phone to handle the full post-meeting workflow — follow-up drafting, LinkedIn content creation, and objection preparation — without needing to open a laptop.

What We Built

Kovil AI designed and delivered a centralised AI web application — the AI Sales & Follow-Up Portal — with three interconnected capabilities, each solving a distinct problem the field sales team faced daily.

01 — Client AI Assistant

Each client account in the portal has its own AI workspace containing the full relationship history: meeting notes (typed or voice-transcribed), objections raised, quotes sent, decision-maker preferences, and deal status. When a rep needs to draft a follow-up after a meeting, they open the client workspace, describe the conversation in a few lines or dictate it by voice, and the AI drafts a personalised follow-up grounded in the complete relationship context — referencing previous conversations, the prospect's stated concerns, and the rep's own tone and style.

Voice input is supported throughout the portal — reps can dictate meeting notes while driving between visits, with the transcript and AI-structured summary waiting for them at their next stop. No notes get lost, and no context gets forgotten.

02 — Content Generation Engine

Field reps in industrial sectors have genuine expertise that resonates with their buyer community on LinkedIn — but they lack the time and writing skill to produce polished content consistently. The Content Generation Engine solves this: a rep records a 60-second voice memo from the field describing a customer problem they solved, a product insight, or an industry observation. The system transcribes it using Whisper and generates 2–3 polished LinkedIn posts optimised for industrial decision-maker audiences, with scheduling and direct publishing via the LinkedIn API.

Posts are generated in the rep's own voice — grounded in their words from the voice memo — not generic corporate copy. This produces content that performs because it is authentic.

03 — Objection Response Database

The portal includes a searchable AI-enhanced library of proven objection responses, built from the client's historical deal data and sales team institutional knowledge. Responses are categorised by objection type (price, timing, incumbent vendor, internal approval process) and by format (email, verbal, text message). When a rep encounters an unfamiliar objection, they search the database and receive 2–3 proven responses, which the AI customises for the specific client context before use.

The library grows with use: when a rep successfully handles a new objection type, they can add it to the database — making the entire team smarter with each deal closed.

Technical Architecture

The portal was built as a progressive web application — installable on iOS and Android home screens, with full voice input support across all tools. The architecture was mobile-first from day one.

  • Stack: Next.js 15, TypeScript, ShadCN component library, and Tailwind CSS — optimised for mobile performance
  • AI layer: Claude and GPT-4o for generation tasks, OpenAI Whisper for voice transcription; model selection varies by task and latency requirement
  • Auth: Secure multi-user login with role-based access — managers can view team activity; reps see only their own client workspaces
  • Integrations: LinkedIn API for direct post scheduling and publishing; CRM webhooks for two-way HubSpot sync; Calendly for meeting booking from within the portal
  • Storage: Supabase for persistent per-client relationship memory — every meeting note, voice transcript, objection, and AI output is stored and fully searchable
  • Deployment: VPS, mobile-first, with voice enabled across all tools

Results

The portal transformed how the sales team operated in the field. Reps who had previously sent follow-up emails two to three days after a meeting — or not at all — were sending personalised, context-aware follow-ups within an hour of leaving the client site, drafted on their phone during the drive.

LinkedIn posting, which had previously happened monthly at best, became a weekly habit because the effort dropped from 45 minutes to under 5 minutes per post. The AI-generated posts — grounded in real field observations — generated more engagement than the polished but generic content the marketing team had been pushing.

One senior AE described it as: "I used to send follow-ups two days after a meeting if I remembered. Now I am sending personalised emails from the car park before I have driven away. The LinkedIn content alone has generated three inbound conversations this month."

Start Your Project

See the engagement model that fits your situation.