The Journal.
Your gateway to AI that actually pays back in marketing: what we build, what it produced, and the parts that did not work.
The 30-day funnel: how a five-person team grew pipeline by 75%
A five-person sales team, a list nobody had time to work, and thirty days. What the system did, what the people kept, and the numbers at the end of it.
Prompt libraries rot. Build evaluation sets instead.
A prompt that worked last quarter is a liability this one. What to keep instead, and how to know when the output has drifted.
Tone of voice is a dataset, not a PDF
Brand books describe a voice. Models need examples of it. How we turn forty pages of guidelines into training data.
Human + AI: where the hand-off should actually happen
We spent a week inside a five-person growth team. The agents did the volume; people kept the judgement. Here is exactly where the line sits — and why it moves.
GDPR and generative AI: a checklist for marketing teams
What has to be true before an agent writes to a European prospect — consent, records, retention and the part most teams miss.
Reading campaign data in hours, not weeks: notes on an analyst agent
By the time a report is assembled, the campaign it describes is already over. How we wired an analyst agent into the media plan — and what it still gets wrong.
U [in] AI featured in AI Odyssey: a founder story about building AI for real B2B growth
It’s always interesting when people outside your own team start asking questions about what you’re building. For us, that’s been happening more often lately.
Browse by what keeps you up at night.
10 articles across 16 topics. The bigger the word, the more we have written about it.
Human + AI: where the hand-off should actually happen
We spent a week inside a five-person growth team. The agents did the volume; people kept the judgement. Here is exactly where the line sits — and why it moves.
Reading campaign data in hours, not weeks: notes on an analyst agent
By the time a report is assembled, the campaign it describes is already over. How we wired an analyst agent into the media plan — and what it still gets wrong.
The 30-day funnel: how a five-person team grew pipeline by 75%
A five-person sales team, a list nobody had time to work, and thirty days. What the system did, what the people kept, and the numbers at the end of it.
Prompt libraries rot. Build evaluation sets instead.
A prompt that worked last quarter is a liability this one. What to keep instead, and how to know when the output has drifted.
Tone of voice is a dataset, not a PDF
Brand books describe a voice. Models need examples of it. How we turn forty pages of guidelines into training data.
- CASE STUDY The 30-day funnel: how a five-person team grew pipeline by 75% 22 SEPT · 1 MIN
- PLAYBOOK Prompt libraries rot. Build evaluation sets instead. 20 SEPT · 1 MIN
- CONTENT Tone of voice is a dataset, not a PDF 19 SEPT · 1 MIN
- PROCESS Human + AI: where the hand-off should actually happen 17 SEPT · 1 MIN
- COMPLIANCE GDPR and generative AI: a checklist for marketing teams 16 SEPT · 1 MIN
- ANALYTICS Reading campaign data in hours, not weeks: notes on an analyst agent 12 SEPT · 1 MIN
- STRATEGY U [in] AI featured in AI Odyssey: a founder story about building AI for real B2B growth 11 SEPT · 4 MIN
Start here.
- 01 U [in] AI founder Olga Bialkovich publishes new article on AI agent consistency in Unite.AI OPINION · 5 MIN
- 02 Why most AI pilots in marketing never reach production STRATEGY · 1 MIN
- 03 Three agents, one lead flow: the math behind −62% CASE STUDY · 1 MIN
- 04 The 30-day funnel: how a five-person team grew pipeline by 75% CASE STUDY · 1 MIN
- 05 Prompt libraries rot. Build evaluation sets instead. PLAYBOOK · 1 MIN
One email every two weeks.
Only what we would read ourselves.
Case numbers, playbooks and the experiments that failed. No digests of other people's news, no launch announcements, nothing you could have found on LinkedIn.
SubscribeNO SPAM · ONE-CLICK UNSUBSCRIBE · GDPR-COMPLIANT