Skip to main content

Outcomes, documented.

Selected client engagements. Names are anonymised where confidentiality is preserved by mandate. Every outcome below is measured and verifiable on request.

CASE 01

Custom LLM Deployment

Professional services firm. Delhi NCR. 80 employees.

The Context

A mid-sized chartered accountancy firm had accumulated eleven years of working knowledge — client correspondence, tax advisory notes, compliance frameworks, internal procedures, and case precedents — spread across shared drives, email archives, and the institutional memory of six senior partners.

When a senior partner retired, the firm lost approximately thirty per cent of its practical knowledge base. New hires took four to six weeks before they could operate independently. Internal queries consumed two to three hours of senior partner time every day.

The Work

We conducted a full knowledge audit, ingested and structured eleven years of documentation, and deployed a Custom LLM on the firm's private infrastructure. The model was trained to answer queries in plain language, cite its source documents, and escalate questions it could not answer with sufficient confidence.

6 weeks4 daysNew hire operational independence

Measured Outcomes

  • New hire operational independence: 6 weeks → 4 days
  • Senior partner time recovered from internal queries: 2.5 hours/day
  • Knowledge base utilisation: from 8% of documents actively referenced to 74%
  • Onboarding cost reduction: 67%
  • ROI timeline: 58 days post-deployment

  • Deployment type: On-premise
  • Timeline: 6 weeks
  • Practice: Custom LLM & On-Premise AI
CASE 02

Website Development

Financial services firm. Mumbai. 35 employees.

The Context

A boutique wealth management firm was operating on a website built in 2019. It loaded in 6.2 seconds on mobile. It had no CMS — every content change required a developer. The consultation request form had a 74% abandonment rate. Their compliance team could not update regulatory disclosures without external help.

The Work

We rebuilt the site on Next.js with a headless CMS, a custom booking integration connected to their calendar system, and a compliance-managed disclosure section their legal team could update independently. The entire engagement ran in parallel with their regulatory audit cycle, requiring careful coordination on content approvals.

3197PageSpeed score

Measured Outcomes

  • PageSpeed score: 31 (mobile) → 97
  • Page load time: 6.2 seconds → 1.1 seconds
  • Consultation request form abandonment: 74% → 31%
  • Inbound consultation requests: +23% in the first 60 days post-launch
  • Editorial independence: full — compliance team updates disclosures in-house
  • Delivery: 18 days from discovery to go-live

  • Stack: Next.js, Sanity CMS, Calendly, Vercel
  • Timeline: 18 days
  • Practice: Website & Application Development
CASE 03

AI Executive Assistant (Slate)

Founder. E-commerce business. Bengaluru. 45 employees.

The Context

A founder running a direct-to-consumer e-commerce business was spending four hours per day on email — reading, triaging, drafting responses, and following up on actions that had been discussed in meetings but not recorded anywhere. She was the bottleneck for every decision that required context she was the only one holding.

The Work

We deployed Slate — our AI Executive Assistant — configured to her communication style, her business context, and her decision-making patterns. The system read and triaged her inbox, prepared her for every meeting with attendee context and agenda summaries, tracked action items from meeting transcripts, and drafted responses in her voice for her review and approval.

4 hours45 minutesDaily time on email

Measured Outcomes

  • Daily time on email: 4 hours → 45 minutes
  • Meeting preparation time: 30 minutes per meeting → 4 minutes
  • Action items missed per week: 6–8 → 0 (tracked automatically)
  • Decisions requiring her direct involvement: reduced by 34% through better context distribution to her team
  • First week of operation: 23 hours recovered

  • Deployment: Private cloud
  • Timeline: 2 weeks
  • Practice: Slate — AI Executive Assistant
CASE 04

Business Diagnostic & Operational Restructure

B2B services firm. Delhi NCR. 37 employees.

The Context

A founder came to us with a problem they could not name. Revenue was growing. The team was capable. But decisions were slow, good hires were underperforming, and the founder was back in the weeds after three years of building a leadership team to avoid exactly that.

Our diagnostic identified three structural failures: decision authority was not mapped below the founder level, two business functions had no clear ownership, and the reporting structure had been built for a twelve-person business and never updated as the team grew to thirty-seven.

The Work

We redesigned the decision structure, introduced role mandates for every senior position, built a decision matrix that removed the founder from fourteen categories of recurring decision, and restructured the reporting cadence.

4.2 days1.1 daysAverage decision cycle time

Measured Outcomes

  • Founder removed from 6 recurring weekly meetings permanently
  • Output in Q2 following restructure: +38% with the same team, no new hires
  • Average decision cycle time: 4.2 days → 1.1 days
  • Two senior hires who had been underperforming began performing within 6 weeks of role mandate clarity
  • Founder working hours: reduced by 11 hours per week

  • Timeline: 8 weeks
  • Practice: Management Consulting & Diagnostics
CASE 05

Web Application Development

Logistics company. Gurugram. 120 employees.

The Context

A mid-sized third-party logistics provider was running their client reporting on a combination of Excel, WhatsApp, and weekly email summaries. Clients had no real-time visibility into shipment status. The operations team was spending three hours per day producing manual reports. Errors in reporting were creating client escalations twice a week.

The Work

We built a custom client portal — a web application giving each client real-time visibility into their shipments, delivery performance, exception alerts, and historical reports. The portal integrated with their existing TMS via API. Clients could download reports, raise queries, and track performance against SLAs without calling the operations team.

3 hours/day20 minutes/dayOperations team reporting time

Measured Outcomes

  • Operations team reporting time: 3 hours/day → 20 minutes/day
  • Client escalations from reporting errors: 2/week → 0 in the first 60 days
  • Client satisfaction score (quarterly survey): 67 → 89 out of 100
  • Portal adoption: 94% of clients active within 30 days of launch
  • One client expanded their contract volume by 40% within 90 days of portal launch, citing visibility as the deciding factor

  • Stack: React, Node.js, PostgreSQL, AWS
  • Timeline: 6 weeks
  • Practice: Website & Application Development
CASE 06

Custom LLM — Retail & E-commerce

Consumer electronics retailer. Pan-India. 600 employees.

The Context

A national consumer electronics retail chain had 600 employees across 34 stores. Product knowledge was inconsistent — sales staff gave different specifications, different compatibility information, and different pricing explanations for the same products. Customer complaints related to misinformation were running at 140 per month. Training cycles took 3 weeks per new hire across every location.

The Work

We built a Custom LLM trained on their complete product catalogue, compatibility matrices, warranty terms, pricing structures, and sales procedures. The model was deployed as a mobile-accessible tool for every sales associate — queryable in Hindi and English. Associates could ask any product question and receive an accurate, sourced answer in under ten seconds.

140/month18/monthCustomer complaints related to misinformation

Measured Outcomes

  • Customer complaints related to misinformation: 140/month → 18/month (87% reduction)
  • New hire training to independent operation: 3 weeks → 4 days
  • Product query accuracy in mystery shopper audits: 61% → 94%
  • Training cost per hire: reduced by 71%
  • Associate confidence rating (internal survey): 54% → 88% within 60 days of deployment

  • Deployment: Private cloud, mobile
  • Languages: Hindi and English
  • Timeline: 8 weeks
  • Practice: Custom LLM
CASE 07

Website Development & SEO

Healthcare services provider. Pune. Private clinic group.

The Context

A group of three private clinics operating under a single brand had no consolidated digital presence. Each clinic had its own separately-built microsite, none of them ranking on Google for their primary service keywords. Appointment booking was handled entirely by phone. New patient acquisition was declining year-on-year despite strong retention numbers.

The Work

We consolidated the three microsites into a single, architecturally unified website with location-specific service pages, a custom appointment booking system integrated with their practice management software, and a full technical SEO rebuild.

0 keywords in top 1034 keywords in top 10 within 90 daysOrganic search visibility

Measured Outcomes

  • Organic search visibility: 0 keywords in top 10 → 34 keywords in top 10 within 90 days
  • Online appointment bookings: 0/month → 140/month within 60 days of launch
  • Phone appointment volume: reduced by 38% (staff time recovered)
  • New patient registrations: +29% in the 90 days following launch
  • PageSpeed score: average 38 across three microsites → 94 on unified site

  • Stack: Next.js, Sanity CMS
  • Timeline: 4 weeks
  • Practice: Website Development
CASE 08

AI Training Programme

Financial services company. Mumbai. 180 employees.

The Context

A mid-sized NBFC had leadership that understood AI was strategically important but a team that did not know how to use it. Twelve tools had been licensed over eighteen months. Adoption was below 20% across all of them. The tools were not the problem. The understanding of how to use them in the context of financial services work was.

The Work

We designed and delivered a four-week AI literacy and application programme for 60 employees across operations, credit, compliance, and client servicing teams. The programme was built around their actual workflows — not generic AI education. Every session included hands-on application to tasks the participants performed every day.

18%71% within 30 daysAI tool adoption post-programme

Measured Outcomes

  • AI tool adoption post-programme: 18% → 71% within 30 days
  • Documented process improvements submitted by participants: 34 in the first month
  • Credit analysis time for standard applications: reduced by 40% by the operations cohort
  • Compliance document review time: reduced by 55% by the compliance cohort
  • Three participants promoted to AI implementation leads within the organisation within 90 days

  • Format: 4-week programme, 60 participants
  • Practice: AI Training & Enablement

Every engagement begins with a defined outcome.

We do not start work until we agree on what success looks like — in specific, measurable terms. That agreement drives every decision from the first day to the last.

If you do not know yet exactly what success looks like for your organisation, that is where we start. Defining it is part of the work.

Start a Conversation →