Skip to content

AI and automation for B2B SaaS operations

I build the systems that let software businesses scale without adding headcount.

Builder, not advisor. Two decades scaling operations inside one company, now designing and shipping production AI agents and automation across customer success and professional services.

Try my live tool: the KPI Positioning Agent →

Lee Grieve

20+ years

Operational scale

Two decades inside one company as it grew from around £3m in revenue to a multi-billion pound valuation, across technical operations, onboarding, deployments and migrations.

Production

AI agents and automation

Real systems shipped and running across customer success and professional services, not slideware. Builder, not advisor.

Live tool

KPI Positioning Agent

A working agent I built at kpi.leegrieve.com, based on Daniel Priestley’s Key Person of Influence framework.

Point of view

Builder, not advisor.

Most people in this space sell decks and frameworks. I ship working systems. The throughline of twenty years has been the same: help a team handle more without simply hiring more, by changing how the work actually flows.

Here is the contrarian bit. Most enterprise AI does not fail because the models are weak. It fails because nobody redesigns the operation around them. A tool bolted onto a broken process just produces the same outcome faster. The win is in the rebuild, not the purchase.

That is the work I care about: agents and automation that take real load off real teams, measured by what they no longer have to do by hand.

Selected work

A few systems I have shipped.

Anonymised where needed. Each one took real, recurring work off a team without adding people to do it.

01

Unified customer success workspace

Problem
A large customer success org worked across three disconnected systems: a CRM, a data warehouse, and a separate revenue-intelligence tool. Before every customer conversation, CSMs assembled the same context by hand, switching tools and copying data. That manual prep did not scale as the book of business grew.
What I built
A single workspace that consolidates all three sources into one surface, shipped to several hundred CSMs in about four months. The context a CSM used to gather manually now loads automatically, in one place.
Outcome
CSMs stopped context-switching across systems for routine prep. The same team carries a growing set of accounts without adding people, because the time that went into finding information now goes into the customer.
02

Decision-support agent for customer guidance

Problem
The highest-value guidance, knowing the next best action for each customer, depended on scarce specialist consultants. It was slow, inconsistent, and impossible to scale by hiring alone.
What I built
A decision-support agent that ingests usage signals, product gaps, account profile and lifecycle context, interprets the patterns, and produces specific next-best-action recommendations. It hands those to a delivery layer that packages them for the customer and to an execution workflow that carries them out. A human stays in the loop on the judgement calls.
Outcome
Guidance that once required a consultant is now generated automatically and consistently. Consulting dependency drops, specialists focus on genuinely complex cases, and the motion scales with the customer base rather than with headcount.
03

Repeatable platform migration system

Problem
Moving a very large customer base onto a new AI platform was traditionally a bespoke, per-account, consultant-led project. Done that way, the migration could never be staffed at the scale required.
What I built
A repeatable, entitlement-governed migration system with two routes by customer size. Smaller accounts move through a light auto-attach motion with minimal manual touch. Larger accounts get a deliberate, guided plan handled by field engineers. My team owns the entitlement layer and the data pipeline underneath both routes.
Outcome
A migration that would have been impossible to resource one account at a time becomes a repeatable motion measured in months, not years. Roughly 100,000 accounts are addressable through the same system, with the smallest moving almost hands-free.

Live tool

The KPI Positioning Agent

An AI agent I designed and built, based on Daniel Priestley’s Key Person of Influence framework. It is a working example of the kind of system I ship: take a proven thinking model and turn it into something people can actually use, end to end.

Built by Lee Grieve. Based on the Key Person of Influence framework by Daniel Priestley, with credit and thanks.

kpi.leegrieve.com

Key Person of Influence, scored live.

Answer a short set of questions and the agent positions you across the five elements of the framework, then tells you where to focus next.

A real, running build. Not a mockup.

Writing

Notes on building with AI in the real world.

All writing

About

From operator to builder.

I joined one software company more than twenty years ago, when it was doing around £3m in revenue. I grew with it through technical operations, customer onboarding, deployments and migrations, as it became a business valued in the billions.

That gave me an unusually long, hands-on view of what actually breaks when a software business scales. The pattern is always the same: the work grows faster than the team, and the instinct is to hire. I spent years finding the other answer, redesigning how the work flows so it does not need more hands.

More recently that has meant AI. I design and ship production agents and automation across customer success and professional services. Now I am productising two decades of operational and build expertise into work and assets I own.

Away from the day job, I write and release original folk-pop music under my own name. Links are in the footer if you are curious.

Get in touch

Building something that needs to scale without scaling the team?

That is the work I do. Send a short message about what you are trying to scale and where it is getting stuck.

Lee Grieve

I build the systems that let software businesses scale without adding headcount.

© 2026 Lee Grieve. Built and shipped by hand.