From one-off research requests to a standing new-business intelligence practice

Client

Date

$500M+ global speciality chemicals manufacturer

08/2026

The challenge

A global speciality chemicals manufacturer makes additives for a mature core market, and finds its growth outside that core with a new business development team of around eight covering markets worldwide. Every question arrived as fresh desk research, commissioned one at a time. Nothing ran on a schedule, and nothing built for one question was reusable for the next.

The solution

FifthRow deployed a new-business intelligence layer inside the team that owns market intelligence and opportunity development. Market questions go in; ranked, source-backed answers come out, with the recurring ones converted into standing systems that run on a schedule, so the same question never has to be researched from scratch twice.

382,406

analyst-hours equivalent, 155 FTEs

~70%

research on a repeatable standard

Repeatable standard298 reusable workflows, 4 themes, ~70% of team activity covered

Research volume1,920 runs, covering 325,116 sources over the past 12 months

Standing, unattended research8 standing systems, 193 unattended runs

Depth per decisionfull market analysis or concept build, on demand

Analyst-hours equivalent382,406 hours saved; equivalent to 155 FTEs

01 — STANDING MARKET RADAR

The team's continuous view of what is moving in its markets. Rather than commissioning a scan when someone remembers to, the radar runs against a fixed set of questions and reports what changed.

Core Activities

Put a named technology or market space under a standing disruption watch for threats and openings, 81 times
Ran a standing market pain-points analysis, first week to last, 73 times
Layered emerging-trend and opportunity scans on top, catching what is starting to happen
Gave a newly prioritised space its own watch, rather than widening the general scan
Structured every output as a source-traceable brief for a business conversation

The Results

A continuously refreshed view of movement in the markets the business is deciding whether to enter, produced to the same standard every time and never dependent on someone having a free afternoon.

232

market reads, none commissioned

02 — TARGET DOSSIER ENGINE

Finding an acquisition or partnership candidate is the easy half. This system takes a candidate and produces the strategic and financial case on it, to the same standard every time.

Core Activities

Identified M&A and partnership candidates to one standard, 63 times against 18,888 sources
Turned each target into a scored strategic and financial dossier, 47 times, 16,844 sources
Added a deep dive past the first screen, technology and IP analysis for early-stage candidates
Extended the same engine into adjacent spaces without rebuilding it
Fed screened-out candidates back into the radar, so nothing is researched twice

The Results

A repeatable, evidence-backed dossier on any acquisition or partnership candidate, so a recommendation reaches leadership with the reasoning and the financial scoring already attached.

35,732

sources behind the target engines

03 — SCHEDULED STANDING SYSTEMS

The systems the team ran most often were taken off the team's hands entirely. Eight of them now run on a schedule with nobody starting them.

Core Activities

Put the highest-frequency systems on a fixed schedule, starting with target identification, running since September 2025
Added an innovation signal system on the same space, and an activity tracker on named players
Set each to compare the current window against the last: what lands is what changed
Extended scheduled coverage into market-trend monitoring and a standing watch on a single company
Sized outputs to be read rather than filed, full source-backed brief underneath

The Results

Standing coverage of the team's priority adjacencies that arrives on cadence, processing 55,459 sources unattended and freeing the team's own attention for the questions that need judgement.

12 months

of unattended standing coverage

04 — SIZING, GO/NO-GO AND CONCEPT BUILD

An opportunity that survives the radar has to be sized and then decided on. This is the part of the practice that produces a number a director will act on.

Core Activities

Sized single ideas and whole domains, 66 runs across two sizing systems
Ran a comprehensive market analysis where the market picture was the decision, 23 times
Added technology analysis where feasibility, not demand, was the constraint
Gave growth ideas a go/no-go verdict, with a TAM-SAM-SOM breakdown where needed
Built the strongest candidate into a full business concept from 3,942 sources

The Results

A sized, evidenced position on any candidate opportunity, produced to a consistent standard, so the go/no-go conversation starts from a number rather than from an argument about the number.

4. SIZING, GO

NO-GO AND CONCEPT BUILD

05 — REGULATORY AND TREND SYSTEMS

Speciality chemistry lives or dies on registration. A set of narrower systems answers the regulatory and trend questions that gate a market entry.

Core Activities

Answered substance notification and formulation registration questions on live products, via two market-specific systems
Ran a regulatory horizon scanner and landscape mapping, 44 runs, showing what is coming, not what has passed
Delivered trend analysis to the team's own brief: two agents, six parallel streams across sustainability, regulatory, technology, product, consumer and market signals
Iterated it across two versions on client feedback, then ran it 13 times over six months
Rebuilt it as a bi-weekly market signals application, live on a call in three minutes: four ranked thought-leadership priorities with signal, audience and point of view

The Results

Registration and regulatory questions answered inside the same practice as the market questions, and a trend capability that moved from a bespoke build to something any user can generate on demand.

3 minutes

from request to a working trend system

What it proved

Over the past twelve months the team ran 1,920 studies against 325,116 sources, 193 of them starting on a schedule. Thirty applications and 298 workflows have run in production, from catalogue templates to a two-agent trend analysis system and a full idea-to-concept build.

The team holds a practice rather than a queue. Four workflow themes carry roughly 70% of research activity: recurring work is repeatable, not rebuilt each time. Eight standing systems cover the priority adjacencies unattended. Across 2,073 runs the platform records 382,406 analyst-hours saved, the equivalent of 155 full-time analysts.

Results

Metric	Benchmark	Result	Improvement
Research runs	manual desk research, no AI tooling	1,920 in the past year	Beyond manual capacity
Sources analysed	limited by the team's reading time	325,116 in the past year	Depth previously unreachable
Scheduled, unattended runs	0 before August 2025	193 across 8 systems	55,459 sources analysed, nobody starting them
Repeatability of the work	every question fresh desk research	Identified 4 themes that cover ~70% of activity	Recurring work now standing systems
Analyst-hours saved	manual desk research, team of eight	382,406 hours, 2,073 runs	Equivalent of adding 155 full-time analysts

SINGLE-OPERATOR REACH

One person who knows the systems deeply serves the whole function, because colleagues route questions in and receive finished output rather than operating anything.

PLUG-IN PLAYBOOKS

The handful of themes the team asks about most run as ready-made systems anyone can fire by naming a company or a topic.

STANDING CADENCE

Systems put on a schedule are still running a year later, so recurring coverage belongs to the practice rather than to anyone's memory.

PURPOSE-BUILT DEPTH

Catalogue systems carry routine scanning, while a decision with money behind it gets a system written for that decision alone.