WEBVTT

00:00:00.000 --> 00:00:01.060
My Own AI Labs

00:00:01.060 --> 00:00:02.780
is soft-launching the AI

00:00:02.780 --> 00:00:04.840
Content Engine. It's a

00:00:04.840 --> 00:00:06.780
customer-run workflow for executives

00:00:06.780 --> 00:00:08.580
and developers, from your

00:00:08.580 --> 00:00:10.200
first brief to reviewed

00:00:10.200 --> 00:00:12.100
content and audience feedback.

00:00:12.560 --> 00:00:13.480
Here's how it works.

00:00:14.708 --> 00:00:18.808
Follow Jordan, an example executive, through the complete content

00:00:18.808 --> 00:00:23.068
journey. Start with Your agents at the top. Choose

00:00:23.068 --> 00:00:26.548
one or several: for example, Codex for setup and

00:00:26.548 --> 00:00:30.208
Claude Code for research and review. Connect and launch

00:00:30.208 --> 00:00:34.268
each selected agent into the same workspace. Their existing

00:00:34.268 --> 00:00:37.588
models and logins stay in place. The harness waits

00:00:37.588 --> 00:00:41.028
for a real tool call before showing an agent

00:00:41.028 --> 00:00:41.648
as verified.

00:00:44.100 --> 00:00:47.600
Jordan advises healthcare leaders and wants a weekly evidence

00:00:47.600 --> 00:00:50.180
briefing, with a short video in their own voice.

00:00:50.880 --> 00:00:55.160
Describe your audience, topics, purpose and preferred channels in

00:00:55.160 --> 00:00:58.960
one page. Choose whether this represents you or your

00:00:58.960 --> 00:01:02.320
organization, and whether you work alone or with reviewers.

00:01:02.880 --> 00:01:06.280
Save the use case. Your connected agents can read

00:01:06.280 --> 00:01:08.980
it, suggest a plan, and ask only for missing

00:01:08.980 --> 00:01:09.460
details.

00:01:11.418 --> 00:01:14.298
Give the publication a name and confirm its audience.

00:01:15.198 --> 00:01:17.638
Your agent can fill these details from your brief.

00:01:18.678 --> 00:01:23.078
Jordan chooses Healthcare AI Brief, for healthcare leaders, covering

00:01:23.078 --> 00:01:26.918
clinical evidence and safe adoption. Refresh to see the

00:01:26.918 --> 00:01:30.638
agent’s saved changes. The agent you work with and

00:01:30.638 --> 00:01:34.078
the writing model are separate choices. Keep a model

00:01:34.078 --> 00:01:36.538
you already use unless you decide to change it.

00:01:37.976 --> 00:01:42.396
Here is how Jordan decides what gets covered. Prioritize

00:01:42.396 --> 00:01:48.056
clinical evidence and practical adoption. Exclude investment speculation. Reject

00:01:48.056 --> 00:01:52.656
unsupported claims about products. Add an editorial rule: prefer

00:01:52.656 --> 00:01:57.716
original research, explain limitations, and skip announcements without evidence.

00:01:58.076 --> 00:02:02.656
Choose trusted feeds, community news, software projects, or connected

00:02:02.656 --> 00:02:07.436
public sources. Save these filters. These filters shape the

00:02:07.436 --> 00:02:09.676
next briefing around your audience’s needs.

00:02:10.716 --> 00:02:14.976
Add the feed addresses of publications you trust. Find

00:02:14.976 --> 00:02:18.216
sources for your chosen topics and inspect the connection

00:02:18.216 --> 00:02:22.196
status. A public data API is another way to

00:02:22.196 --> 00:02:25.556
receive articles or records. Your agent can help with

00:02:25.556 --> 00:02:29.576
a supported read-only connection that returns headlines and original

00:02:29.576 --> 00:02:32.956
links. A catalog listing is not proof it works.

00:02:33.956 --> 00:02:38.736
Private systems, sign-in-only sources, PDFs and arbitrary websites need

00:02:38.736 --> 00:02:39.776
a separate integration.

00:02:41.407 --> 00:02:45.287
Responsible AI is built into this workflow. The harness

00:02:45.287 --> 00:02:49.087
checks original sources and keeps their links, so you

00:02:49.087 --> 00:02:52.807
can inspect the evidence. Story visuals carry their evidence

00:02:53.077 --> 00:02:56.947
and are reused across the newsletter and video. Review

00:02:56.947 --> 00:03:01.767
includes visual checks and transcript matching. Failed or unavailable

00:03:01.767 --> 00:03:06.207
checks remain visible. These controls reduce errors; they do

00:03:06.207 --> 00:03:10.267
not guarantee every claim. Human review, consent for your

00:03:10.267 --> 00:03:13.847
voice or likeness, and fresh approval for changed content

00:03:13.847 --> 00:03:14.727
remain essential.

00:03:16.259 --> 00:03:20.299
The first voice choice is local Voicebox. Record a

00:03:20.299 --> 00:03:23.179
clear sample of your own voice, create a named

00:03:23.179 --> 00:03:26.899
profile, and keep the local studio running. Find your

00:03:26.899 --> 00:03:30.999
local voices and select your authorized profile. This walkthrough

00:03:30.999 --> 00:03:35.079
also uses locally generated narration. If the chosen voice

00:03:35.079 --> 00:03:39.639
fails, the job holds rather than switching identity. A

00:03:39.639 --> 00:03:42.459
built-in narrator is available if you prefer a quick

00:03:42.459 --> 00:03:43.159
first preview.

00:03:44.658 --> 00:03:47.858
Now see the finished product. These are three real

00:03:47.858 --> 00:03:51.638
renders of the same short briefing: narrated slides, presenter

00:03:51.638 --> 00:03:56.538
only, and presenter alongside slides. The presenter is fictional

00:03:56.538 --> 00:04:00.258
and AI generated. Narrated slides are available in Free.

00:04:01.258 --> 00:04:04.918
Presenter production is a Pro option, with setup handled

00:04:04.918 --> 00:04:08.558
behind the scenes by MyOwnAI Labs. You choose the

00:04:08.558 --> 00:04:13.258
result, without navigating animation vendors. Watch the same content

00:04:13.258 --> 00:04:14.158
in each layout.

00:04:15.190 --> 00:04:18.770
start with reliable sources. Check each claim,

00:04:19.330 --> 00:04:21.830
review the finished video, and publish only

00:04:21.830 --> 00:04:24.750
when you are ready. Then use audience

00:04:24.750 --> 00:04:26.890
feedback to improve your next briefing.

00:04:29.051 --> 00:04:33.295
Jordan selects LinkedIn and YouTube together. You can choose

00:04:33.295 --> 00:04:36.091
as many channels as you need, then connect each

00:04:36.091 --> 00:04:40.355
account separately when ready. Google owns YouTube, so a

00:04:40.355 --> 00:04:45.511
Google application ID identifies the application requesting upload permission.

00:04:45.729 --> 00:04:49.160
It is not your email address. Your agent can

00:04:49.160 --> 00:04:52.909
guide that setup; you enter private credentials and complete

00:04:52.909 --> 00:04:56.419
the service’s own consent screen. You can make a

00:04:56.419 --> 00:04:59.671
private preview before connecting any publishing account.

00:05:00.554 --> 00:05:03.826
Your draft is a private first version: a sourced

00:05:03.826 --> 00:05:07.197
newsletter and a narrated video about the same story.

00:05:07.435 --> 00:05:10.905
Prepare the tools and check your writing model, then

00:05:10.905 --> 00:05:15.168
authorize generation. Nothing is published by that button. For

00:05:15.168 --> 00:05:19.035
this walkthrough we use a labelled teaching package with

00:05:19.035 --> 00:05:22.723
the finished sample video, rather than generating a factual

00:05:22.723 --> 00:05:26.748
healthcare story. In your workspace, inspect the actual sources,

00:05:27.165 --> 00:05:29.465
words, narration and visual evidence.

00:05:30.317 --> 00:05:33.317
Preview is the beginning of review, not the end

00:05:33.317 --> 00:05:36.137
of the engine. Watch the video and read the

00:05:36.137 --> 00:05:40.837
companion newsletter. Approve only that exact version. A changed

00:05:40.837 --> 00:05:45.137
package needs fresh approval; team workspaces can require another

00:05:45.137 --> 00:05:49.357
reviewer. The next controls publish the newsletter, post the

00:05:49.357 --> 00:05:52.497
approved video to a selected channel, and add its

00:05:52.497 --> 00:05:55.997
link back to the newsletter. Each external action needs

00:05:55.997 --> 00:05:59.257
its own explicit browser click. We do not publish

00:05:59.257 --> 00:06:00.117
this demonstration.

00:06:01.691 --> 00:06:05.751
After publishing, open Results. Delivery tells you whether a

00:06:05.751 --> 00:06:09.811
channel accepted the post. Independent checks confirm the published

00:06:09.811 --> 00:06:14.471
post. An uncertain attempt needs inspection, not an automatic

00:06:14.471 --> 00:06:18.951
retry. Audience analytics show views and engagement. They also

00:06:18.951 --> 00:06:22.971
show which channels responded. The numbers shown here are

00:06:22.971 --> 00:06:28.391
clearly labelled training data. Missing measurements remain unknown. Collect

00:06:28.391 --> 00:06:32.511
fresh metrics from supported, authorized accounts before making decisions.

00:06:33.811 --> 00:06:38.659
The engine continues into audience follow-up. Questions and corrections

00:06:38.659 --> 00:06:41.450
rise to the top of the inbox. Review a

00:06:41.450 --> 00:06:45.487
viewer’s response, prepare a helpful reply, and approve its

00:06:45.487 --> 00:06:49.029
exact wording before sending. Your agent can suggest a

00:06:49.029 --> 00:06:53.106
draft; it cannot grant approval. X and YouTube support

00:06:53.106 --> 00:06:57.974
account-authorized collection and replies. Other channels use manual capture

00:06:57.974 --> 00:07:02.802
and posting handoff. Reactions help identify interest, but do

00:07:02.802 --> 00:07:04.187
not create reply threads.

00:07:05.122 --> 00:07:09.868
The cost dashboard shows recorded writing-model calls, reported dollar

00:07:09.868 --> 00:07:13.958
amounts, and calls whose prices were not returned. These

00:07:13.958 --> 00:07:17.930
example entries are training data, not your bill. A

00:07:17.930 --> 00:07:20.293
missing price does not mean a service is free.

00:07:20.491 --> 00:07:24.681
This dashboard does not yet track agent subscriptions, conversation,

00:07:25.038 --> 00:07:29.208
voice, avatar, rendering or tunnel fees. Check those account

00:07:29.208 --> 00:07:32.425
statements for the full cost. Use this view to

00:07:32.425 --> 00:07:35.165
understand recorded usage and its limits.

00:07:36.061 --> 00:07:39.141
Use the results and audience questions to refine your

00:07:39.141 --> 00:07:43.481
next edition. Update the source filters, ask your agents

00:07:43.481 --> 00:07:46.921
to investigate the strongest questions, and prepare a new

00:07:46.921 --> 00:07:50.221
private draft. Talk to your harness is an optional

00:07:50.221 --> 00:07:55.161
conversation space with microphone controls and a transcript. You

00:07:55.161 --> 00:07:58.821
can interrupt or type during a connected session. Its

00:07:58.821 --> 00:08:03.521
voice connection is separate from publication narration. Final approvals

00:08:03.521 --> 00:08:05.581
and external actions stay in the browser.

00:08:07.127 --> 00:08:09.207
Book a demo at My Own AI Labs dot

00:08:09.207 --> 00:08:12.827
com. Bring your use case, and we’ll estimate the

00:08:12.827 --> 00:08:16.187
hours you could save each week. The open-source engine

00:08:16.187 --> 00:08:19.647
is free. Pro is planned at ninety-nine dollars a

00:08:19.647 --> 00:08:24.287
month for premium workflows and guided support. Agent subscriptions

00:08:24.287 --> 00:08:25.927
and usage are separate.

00:08:26.583 --> 00:08:27.823
Explore the AI Content

00:08:27.823 --> 00:08:29.163
Engine at my own

00:08:29.163 --> 00:08:30.323
AI labs dot com.

00:08:31.063 --> 00:08:32.383
Free beta access will

00:08:32.383 --> 00:08:34.323
be by invitation. If

00:08:34.323 --> 00:08:35.903
you're interested, request a

00:08:35.903 --> 00:08:36.163
demo.