02 June 2023

Photoshop’s New A.I. Update + 4 More A.I. Tools You Must Try (#125)

1Frameworks

10Insights

Frameworks in this episode

Leadership5 steps

Possibility-First Technology Training

Teach teams what technology can achieve before teaching individual procedures

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster13:30

Why Acceptable AI Writing Is Not Enough for LinkedIn

Hushl can generate competent LinkedIn content, but the hosts question whether that output is meaningfully better than a general-purpose model. Their larger criticism is that average writing will not stand out on a competitive channel, even when a tool is trained or tuned for that channel.

Hushl produces a usable first draft from a simple brief.
Generic trend posts may earn little engagement.
Channel-specific tooling does not guarantee exceptional writing.
AI drafts still require distinct ideas and strong editing.

“The writing is okay, but, actually, okay is not good enough on any single channel at all in terms of writing.”
Kieran Flanagan · 14:30

“You actually have to be truly great.”
Kieran Flanagan · 14:30

#linkedin#content writing#hushl#quality

Hot Take· 2

Hot Take09:30

Every Software Product Will Get a Natural-Language Copilot

The Windows 11 copilot prompts a broader prediction that natural language will become the standard access layer for software. If users can request outcomes conversationally, specialist administration roles and conventional interface advantages may face disruption.

Copilots will execute tasks inside individual software products.
Natural language lowers the barrier to using complex features.
Specialist software administration may become less necessary.
The interface shifts from navigating controls to expressing intent.

“every software product is going to have a co-pilot, and that co-pilot will do all of the things that you want to do in that…”
Kieran Flanagan · 10:00

“On top of all software, there is gonna be a natural language layer.”
Kieran Flanagan · 10:00

#copilots#natural language#software#windows

Hot Take11:30

AI Could Erase Ease of Use as a Product Differentiator

If copilots make complex software accessible through conversation, ease of use may cease to distinguish competing products. Brands that rely on simplicity as a core promise may need stronger differentiation based on outcomes, data, trust, or unique capabilities.

Natural-language interfaces can flatten usability differences.
Easy-to-use may weaken as a defensible product benefit.
Software brands may need to revise their positioning.
Underlying value becomes more important when operation is commoditized.

“what happens to every brand's strap line when ease-of-use is commoditized across software.”
Kieran Flanagan · 11:30

“easy-to-use will probably go away as a product benefit.”
Kipp Bodnar · 11:30

#positioning#product strategy#usability#ai

Explainer· 2

Explainer00:30

Why Photoshop's Generative Fill Changes Image Editing

Adobe brought Firefly's generative capabilities directly into Photoshop, allowing users to expand, modify, and reconstruct images through natural-language requests. The hosts argue that integrating AI with established editing tools gives Adobe an advantage over standalone generators when users need production-ready assets.

Generative Fill can expand an existing image beyond its original frame.
Users can add or alter image elements with natural language.
Photoshop supplies the editing controls needed to finish generated assets.
Established software companies can distribute AI through existing workflows.

“the ability to like edit the images through natural language is just absolutely mind-blowing.”
Kieran Flanagan · 02:30

“you have no restrictions anymore in terms of like how you can take imagery and tailor it to any of your needs.”
Kieran Flanagan · 02:30

#adobe#firefly#photoshop#image generation

Explainer06:30

Google and TikTok Are Turning Short Video Into a Search Surface

Google is making YouTube Shorts more searchable while TikTok is testing Tako, a chatbot that recommends videos in response to questions. Both moves anticipate rapidly growing video supply and aim to improve discovery, relevance, and platform monetization.

Google is applying AI to YouTube Shorts discovery.
TikTok's Tako can recommend videos based on conversational questions.
Food, workflows, and other visual categories may compete with text search.
More searchable video gives businesses additional opportunities to be discovered.

“How do I leverage these innovations to help people better discover my business through short form video?”
Kipp Bodnar · 07:30

“TikTok really is a potential competitor, not for the entirety of Google--”
Kieran Flanagan · 08:30

#tiktok#youtube shorts#video search#discovery

Tool· 3

Tool05:00

Paragraphica Turns Location Data Into an Imagined Photograph

Paragraphica uses location information to describe the user's surroundings and generate a synthetic image of the scene. Rather than recording the street optically, it converts contextual data into a prompt and produces an idealized interpretation.

The device builds prompts from location data.
It generates rather than directly captures the scene.
Generated results can remove clutter and alter lighting.
The experiment demonstrates a new relationship between cameras and image models.

“it's a camera that takes photos using your location data, describes the place you are at, and then converts it into an AI-generated image.”
Kipp Bodnar · 05:00

#paragraphica#camera#location data#generative ai

Tool12:00

Vimcal Uses Natural Language to Simplify Complex Scheduling

Vimcal applies natural-language interaction to calendar coordination, including identifying suitable times and drafting availability text for email. The hosts see the greatest value for people managing complex schedules across many groups.

Vimcal identifies scheduling options conversationally.
It drafts availability text that can be pasted into email.
The product extends AI into a familiar productivity workflow.
Complex calendars offer the clearest use case.

“it's like superhuman for like calendar.”
Kieran Flanagan · 13:00

“If you're somebody who's juggling a ton of different meetings between a lot of different groups of people, Vimcal seems pretty awesome.”
Kipp Bodnar · 13:00

#vimcal#calendar#productivity#scheduling

Tool15:30

MagicForm Turns Company Knowledge Into an AI Sales Concierge

MagicForm scans website pages and uploaded documents, then uses that information to train a chatbot for sales enablement and customer questions. The hosts view this as part of a broader shift toward always-available AI concierges across go-to-market operations.

The tool ingests website pages and internal documents.
It can accelerate sales-representative product training.
A chatbot can answer company questions continuously.
AI may commoditize fast, knowledgeable customer support.

“with one line of code, you have a chat bot trained on all your data instantly.”
Kieran Flanagan · 16:00

“it is gonna be such a table stakes experience to have an AI concierge across your entire go to market.”
Kieran Flanagan · 16:00

#magicform#sales enablement#chatbot#customer experience

Takeaway· 2

Takeaway03:00

The Winning Creative Stack May Use More Than One AI Tool

The hosts distinguish initial image generation from the manipulation required to create a finished business asset. They expect specialized applications to form a new creative suite rather than one universal AI product handling every stage equally well.

Midjourney and DALL-E are useful for creating initial images.
Firefly is positioned as stronger for manipulation and finishing.
Website and newsletter assets often require format-specific edits.
Creative workflows may combine several specialized AI products.

“There's not just gonna be one magical AI application, I think, that does all of this stuff, Kieran.”
Kipp Bodnar · 03:30

#creative tools#midjourney#workflow#design

Takeaway17:30

The Next CRM Serves Buyers as Well as Sellers

The hosts argue that CRM data should power customer-facing assistance as well as internal sales workflows. A shared AI layer can help buyers get answers at any time while also training representatives and supporting the complex portions of a sale.

CRM systems have historically centered on sales representatives.
AI makes company data directly useful to buyers.
The same knowledge layer can support customers and train sales teams.
Better access to information should reduce friction in purchasing.

“I think we are at the point where every CRM is for both buyers and sellers.”
Kipp Bodnar · 17:30

“You give people the ability to buy in any way they wanna buy.”
Kieran Flanagan · 18:00

#crm#buyers#sales#customer experience