Practical Consulting

Selected work

Anonymised, but real.

Most clients prefer not to be named. The engagements below are real in pattern — sectors, problems, and outcomes are accurate; identifying details aren't. Composites where the underlying story has been told more than once.

SMB telco

Voice infrastructure modernisation

Lifting a 12-year-old voice platform without dropping a customer

A regional telco running a multi-tenant voice platform for thousands of downstream business seats had hit the ceiling on its on-prem stack. Vendor support windows were narrowing, capacity expansion was getting expensive, and the platform couldn't ship features competitors were starting to give away as standard.

Migration ran in three phases over four months: customer-by-customer routing redesign, parallel-platform cutover with number-portability handling, and decommission of the legacy stack. The trickiest part wasn't the new platform — it was making sure no active call dropped during the swap.

Outcomes

  • Zero disruption to in-progress calls during cutover
  • ~40% reduction in ongoing platform OPEX
  • Call recording and basic analytics shipped to all tenants within 60 days

SMB telco

AI voice agent rollout

An AI agent on the front line of customer support

An SMB telco was fielding rising volumes of routine support traffic — billing questions, outage checks, service activation — with a small team that couldn't keep wait times under four minutes during peaks. They didn't want a chatbot bolted to a phone number; they wanted a voice agent that integrated with their actual systems.

Deployed an LLM-backed voice agent connected to the telco's billing platform and outage status API, with strict scope. The agent could answer questions, run safe self-service actions, and escalate cleanly to a human at any point. Carrier-grade voice path, real transcription, real call quality.

Outcomes

  • ~55% of inbound calls fully resolved by the agent without human handoff
  • Average wait time for escalated calls dropped to under 60 seconds
  • Support team capacity recovered without new hires

Professional services firm

AI voice agent · after-hours intake

A receptionist that knows when not to talk

A mid-sized professional services firm was losing leads to after-hours and overflow calls. Partners didn't want a 24/7 answering service, and manual triage of voicemail was costing intake staff most of every Monday morning.

Voice agent deployed to take new-enquiry calls outside business hours, capture caller intent and contact details, give realistic next-step expectations, and hand everything cleanly to the firm's intake system. Hard guardrail: the agent never gave advice or made commitments — uncertain calls were always flagged for a human callback first thing next morning.

Outcomes

  • ~40% more after-hours enquiries captured vs voicemail-only baseline
  • Zero false promises (the agent simply could not make any)
  • Higher conversion of after-hours enquiries to billable engagements

Got a problem like one of these?

Or a different one entirely. The first conversation is free.

Get in touch →