NORLA FORUM
Open discussions between the operators, builders, instructors and sellers on Norla. Anyone can read along without an account; sign in to start a topic or reply.
Welcome. This forum is for the people who run marketing work with agents: operators, builders, instructors and the studios publishing in Extensions. Three rules. 1. Keep client na…
In Orchestrator the reviewer returns the first draft, then approves the revision almost every time. Is it reading the revision properly or just happy to see a change?
We have run our docs announcements on Orchestrator since July. This week I rebuilt the whole docs pipeline as a new workspace: researcher, brief writer, copywriter, reviewer. Three…
A colleague asked whether to start the trial or go straight to a plan. I have been on Pro since May and realised I could not explain it in one sentence. Trial is fourteen days once…
We have used the launch content workspace internally for months, on the team account. This week we ran it live in a client meeting for the first time. Researcher, brief writer, cop…
Finished the six-week pace. Honest retro. Kept: the workflow brief (module 2), the evaluation floor (module 3). Dropped: the module 1 reading, I had read it before. Surprised by:…
I bought the onboarding agent four weeks ago. Used: the segment logic, the claim-check gate, the metrics sheet. Not used: half the prompts, because our product has fewer features t…
The note on approvals having a scope and an expiry is clear in principle. What I would find useful is a worked example: an approval that expired, what the agent did next, and what…
Two agency buyers asked for the receipt to show their company name and address. The receipt shows the account holder. Is there a way to do this on Norla, or do people just send a s…
the caption agent is fine in the second language now. the hook agent is not: english hooks are questions, german hooks that are questions sound like a quiz show. translating the ho…
Following up on the earlier thread about shadow runs: what is in shadow mode on your side this month, and what would it take to switch it on?
Proposing a reading group on the editorial note about experiments not being finished until the decision is recorded. One week, one note, then a thread of what we each changed. Who…
Looked at a treatment yesterday and could not remember which paragraphs I wrote. Not sure if that is good or bad.
The pipeline will happily produce twenty clips from a one-hour recording. Six feel right; twenty feels like spam. Is there a number people have settled on, or is it a judgement cal…
The 'My discussions' filter shows threads I started. I want the ones I replied to.
Switched the copy agent to a newer model for cost reasons. Factual accuracy went up, tone went flat. Everything in the system prompt was the same. Which parts of a voice guide surv…
Same as the check-in but for the things that went right. One line each, no humility required.
Buyers of the brand kit builder keep asking what the output looks like. I describe it; they want to see it. Would a sample output on the listing help, and how do you show one witho…
If you are reading this you are one of the first people here, so the forum is small and that is fine. When you have a minute, reply with three things: what you do, one job an agen…
No posts since Monday. Budgets locked, or just busy?
The recap email agent opens with 'Apologies for the delay' even when there was no delay. The instructions say nothing about apologies. Where does this come from and how do I stop i…
Sent? Opened? Clicked? Customer did the thing? Our agent marks 'done' when the email is approved, which is about four steps before anything happens. Where do you put the line?
Ranges only. Mine: units up versus September, revenue flat because the cheap packaging copy system is doing the volume.
I did, mostly because the 10% fee line made me curious about the rest. The payout timing (14 days after delivery) was the part I had not expected. Did anyone else read it, and was…
Office hours are in course-clinic but half the people who would benefit are in operator-lounge and never look at the clinic. Cross-posting feels noisy. What is the norm?
We run short-form content for three brands with a repurposing pipeline and a brief agent. Looking for someone 2–3 days a week to own the review step: read drafts against the brief,…
Listings are text only. A sample output image would answer most of my pre-sale questions.
I will be there for the field marketing track. If a few of us are going it would be good to put faces to handles. Coffee the morning of day two?
I have sixty onboarding emails with reviewer scores and no time to turn them into a proper evaluation set. If someone wants to pair on it for an afternoon, I will share the data (r…
The editorial pages let you keep private notes on this device. I have been using it as a scratchpad for what I would test. Curious whether others use it or whether it is just me.
One line each. Me: the lifecycle agent is rewriting our onboarding sequence and I am reading every draft.
Small one. Export on /orders gives dates like 06/10/2026. My spreadsheet is set to US and reads that as June 10. Could the export use ISO dates (2026-10-06)? Everything else in the…
The launch brief agent now drafts release notes from the facts sheet. Product says they are the clearest notes we have shipped. Support says customers quote them back when somethin…
We have an agent that flags fatigue when frequency rises and CTR falls for three days. It is right about half the time. The other half it is a weekend or a budget change. Is anyone…
After moving the rubric into the cached prefix, scores got more consistent. Too consistent. We think the model stopped re-reading the rubric for edge cases because the prefix is tr…
Office hours at 4pm UTC is midnight for me. Not a complaint, just asking whether there are enough of us in APAC for a second slot or a room.
We run research → brief → draft → repurpose with four agents. Right now a human reviews only the final draft, and by then the brief is usually where the problem started. Where do…
Small win. The research-to-brief contract requires an acceptance owner. A brief came through with none, the pipeline stopped, and the person who would have reviewed it was on holid…
Client review tomorrow. The creative scoring agent ranked assets; a human chose the test; a human wrote the final copy. The client will assume the agent did everything or nothing.…
Listed a brand kit builder last month. Two refund requests, both from people who expected it to design the logo. The listing says it does not. Nobody read it. What do you put in t…
Evaluation sessions come in paid. I message the buyer to schedule, start the work when they reply. Two buyers took a week to answer and the order sat in Confirmed. Should I mark it…
I joined to evaluate platforms for a forty-person team and ended up buying an evaluation session instead of a platform. If you joined recently, say what you came for and what you a…
Listed the photography direction session with 'remote' in the format line and 'in person optional' in the description. A buyer read the description, expected me in Madrid, and I wa…
Field marketing here. After every partner call I have a forty-minute transcript and I need a one-page brief by the next morning. I keep writing the prompt from scratch. Does anyone…
The paid-media room asks for numbers as ranges if the client would mind. Half the posts still give exact CPAs. Is the range rule useful, or should the room just be exact and let pe…
Show notes for a two-hour episode: the first hour is summarised well, the second hour gets vague. Same model, same prompt. Is this a known thing and what do people do about it?
Thirty years in marketing, two in agents. Half the agents I review are a checklist with a chat interface in front of it. The chat interface adds cost, variance and a false sense of…
Evaluating the Content Engine Suite template (installed in Orchestrator) for a 40-person team. Our engineers say they can build the same thing in a sprint. They always say that. Ha…
The repurposing pipeline now produces thirty clips a week. Brand review was built for three. The reviewer skims, signs everything, and then a clip with the old logo goes out. How d…
Three months after the confident-wrong report. Every number in the weekly now has a footnote with the table and the date range. The agent refuses to write a number it cannot cite.…
Thinking about a second listing: the voice guide as a prompt pack with examples per channel. Buyers so far are individuals, but two agencies asked whether they can use it across a…
The onboarding sequence sends on a timer relative to signup. A customer in Auckland got a "good morning" email at 2am. Do you let the agent infer timezone from signup data, or hold…
Module 3 asks for an evaluation plan before widening scope. Mine: ten briefs, a reviewer scores each 1 to 5 on completeness and sourcing, ship if the average is above 4. Instructor…
It attributed a revenue lift to an email campaign that had not sent yet. The chart was beautiful. Nobody checked the date range. What guardrails do you put on narration agents so…
Not asking for the best. Asking what you use day to day for podcast and call transcripts, and the one thing that annoys you about it.
We run twelve agents in the editorial pipeline and the change log is a Slack channel nobody scrolls back through. When a draft comes out wrong, the first question is always "did an…
course-clinic office hours Thursday 4pm UTC. This week's theme is the failure drill from the reliability field note: bring one run that went wrong and we will work out what an oper…
None of them are about AI. 1. A book on technical writing, for the sentence 'the reader has no obligation to understand you'. 2. A book on checklists in surgery, for the differenc…
Opened Pearly Lifecycle in Extensions. First listing is the 14-day onboarding sequence we run in production, with the claim-check gate from the thread above. Happy to answer scope…
The editorial note on budgeting for accepted work says to count retries. We started doing it and the real cost per accepted brief is 2.4 times the naive per-call cost. Anyone have…
Listing an editorial pass as a service. The buyer pays first, then we agree a schedule in messages. What do you put in the listing so the schedule conversation is short?
A support-adjacent agent drafted a reply promising a refund window we do not offer. It was caught in review, but it was confident and well written. What is the pattern for keeping…
Our experiments dashboard is full of results and empty of decisions. Six months later nobody remembers why we stopped the pricing test. Is anyone using an agent to write the decisi…
Switching the video pipeline to cached system prompts cut cost per clip by about a third. It also made us restructure prompts so the stable part comes first. Curious what other str…
Our scoring agent ranks a static ad low on hook clarity. The buyer says it is the best performer in the account. The rubric is calibrated on forty examples. Do we trust the rubric,…
Bought AI Agents for Growth Marketers. Six weeks is the suggested pace but I have one weekend before a launch. Which module gives the most immediately useful thing?
Opened a studio last week with an evaluation kit. Three orders so far and two buyers messaged me questions that are answered in the description. Do people skip the description and…
We have a documentation agent that proposes edits to the design system docs. It has been in shadow mode for three weeks: it writes to a branch, a human merges. The merge rate is no…
Rewriting our voice guide so the agents can actually follow it. The old one was adjectives (warm, confident, clear). The new one is twelve example sentences per situation. Which wo…
The Creative Intelligence template gives every ad concept a fatigue mark before we spend. The scores correlate with results about 70% of the time. Is that good enough to kill conce…
Short review of Data-Driven Campaigns with AI: the module on reading experiment results is the best written explanation of significance I have seen for marketers. I now use the wor…
I keep a set of 40 real briefs with known-good outputs and run every prompt change against them. It catches regressions but it is slow to maintain. What does everyone else do — eva…
After every partner event we used to spend a day on recap emails. Now our lifecycle workspace in Orchestrator drafts one per attendee segment from the event notes, and the field te…
I build the marketing data models. Since spring, half of what I build is read by agents before any person sees it. Three things I did not expect: 1. Column names matter more than…
Clients now send me agent-written feedback on rough cuts. It is polite and thorough and almost unusable: no timecodes, lots of 'consider tightening the middle section'. Editors, w…
August is quiet for a lot of you, which makes it a good month to fix the small thing that has annoyed you since spring. Reply with one fix, however small.
Product photographer. After a shoot I write notes for the retoucher: remove dust, keep the texture on the leather, match the white to the last batch. It takes an hour per shoot and…
Customer marketing lead. Three of us are considering the review course together rather than one person doing it and reporting back. The argument for: we would all use the same voc…
Our docs agent keeps component pages in sync with the code, and the text is genuinely good now. The screenshots are another matter. Half of them show a version of the button from l…
I have settled into a split that works for me. The agent collects evidence: reviews, competitor claims, search language. It sorts that into themes with sources. I write the argumen…
I build the email templates the marketers fill. Since the writing agents arrived, the copy arrives faster and breaks the template more often: a headline that is three lines long, a…
Growth engineer, slightly ranty. Three teams this year asked me for an agent to 'find growth opportunities'. All three had tracking where half the events fired twice and the other…
Our planning agent fills the calendar from a topic list. It keeps scheduling the same topic two weeks apart with a slightly different title, so it looks varied and is not. We trie…
Bootstrapped analytics tool. I am the marketing team. Here is the current split, since people keep asking. Agents: first drafts of every blog post, the newsletter outline, the soc…
I set up agents for clients and then my contract ends. Twice now I have come back six months later to find the agents still running on prompts nobody understands, or switched off b…
We gave an agent last quarter's results, the targets and our backlog, and asked it for a plan. It produced a very tidy plan that spread effort evenly across everything, which is th…
Running paid social in three markets. I have tried three agents from Extensions for weekly reviews. The difference between them is not the writing; it is whether they read the expo…
I will start. Apology emails, anything legal-adjacent, and the first line of a launch announcement. The first two are risk. The third is taste, and I am not sure I can defend it.…
I make short films for brands. I tried letting an agent turn the client brief into a first storyboard (text frames, not images). It is fast. The frames are all competent and none o…
Our change-summary agent needs to know the token names. Pasting the whole list into the instructions works until someone renames a token and forgets the prompt. Has anyone moved t…
Our review process did not change when agents arrived; we just stopped being honest about it. A junior's first draft gets a conversation. An agent's first draft got a shrug and a r…
Three patterns after about forty episodes: 1. It turns jokes into claims. A guest says 'we basically invented email' as a joke and it becomes a timestamped highlight. 2. It over-t…
Our nurture sequences are now mostly agent-drafted, human-edited. Last quarter the pipeline number from nurture came in under target and the first question in the review was 'was t…
We are drafting win-back emails with an agent. The more history we give it, the more personal the email gets, and past a point it starts to feel like surveillance. 'We noticed you…
We let an agent write the subject line variants for 200 campaign sends since November. Numbers, not opinions: - Open rate across all sends: flat against the human-written baselin…
We used to call things 'the GPT one' and 'the new one'. Now every agent is named after its job: Report Drafter, Inbox Router, Notes Cleaner. When we swap the model underneath, nobo…