Wouldn’t it’s implausible when you discovered the antidote to writers block. Or when you may brainstorm lots of of concepts instantly.

It seems like a reasonably cool superpower proper?

It’s one I’ve typically wished for, and one 1000’s of Buffer prospects have hoped for.

It’s additionally a dream shared by Buffer’s Senior Product Supervisor Diego Sanchez.

Which is why, Diego was (and nonetheless is) so excited concerning the energy of synthetic intelligence to not solely ease these burdens but additionally revolutionize your complete method to social media.

Again in February, Diego helped launch Buffer’s AI Assistant.

On this weblog submit, we’ll take you behind the scenes at Buffer, strolling by means of the start of Buffer’s AI, exploring its improvement course of, its integration with OpenAI, and the important function person suggestions performs in shaping its future.

Why AI?

Diego, a Product Supervisor (PM) at Buffer, had been eager about AI for years and has labored within the area for even longer.

Beforehand a PM at Typeform, Diego labored with machine studying engineers on AI functions in 2017.

When he joined Buffer in 2019, he was eager to discover AI functions throughout Buffer’s merchandise, reminiscent of what was previously Reply and Have interaction, however are actually engagement options inside Buffer.

After realizing the potential of AI in content material era, Diego introduced his concepts to Maria, Buffer’s Chief Product Workplace, and Joel, Buffer’s Founder, and CEO. They gave him the inexperienced gentle to go forward.

Working carefully with Dinos (a product engineer on Diego’s workforce), they began exploring get to market shortly. Because the market demand for AI options grew, Buffer accelerated its efforts, and Buffer’s AI Assistant began to take kind.

Figuring out key ache factors

Diego’s workforce labored carefully with customers in our beta neighborhood to establish friction factors and alternatives for optimization throughout the product.

They discovered that small companies and creators typically wrestle with developing with submit concepts, discovering their model’s tone and voice, and managing content material for a number of social media channels.

The problem of making a excessive quantity of participating content material was notably fascinating.

To assist customers overcome this problem, Diego regarded to use AI expertise to content material era.

Improvement course of and Open AI

Buffer’s AI Assistant improvement course of started by fixing key friction factors within the product.

The workforce determined to concentrate on text-to-text functions, aiming to assist customers write captions, generate submit concepts, rewrite content material in numerous tones, and broaden on concepts.

Diego and Dinos explored potential options envisioning ideas reminiscent of AI-assisted hashtag strategies, content material concept mills, and extra.

Buffer’s AI Assistant integrates with OpenAI by way of its API, permitting it to doubtlessly work with different providers sooner or later. This flexibility permits Buffer to check totally different providers, supply personalised experiences, and probably develop in-house AI options.

Diego picked OpenAI on account of its highly effective APIs, varied fashions, and fixed updates.

Its cheaper and quicker fashions are perfect for clearly outlined issues, whereas extra highly effective fashions like DaVinci or ChatGPT4 can deal with open-ended functions.

Now with a working prototype and a strong AI API plugged in, Diego wanted extra person suggestions.

Consumer suggestions performed a vital function in shaping Buffer AI’s performance. It helped the workforce make changes and establish new alternatives and use circumstances.

For instance, one early adopter steered an AI-assisted Twitter thread feature, which Buffer is now actively exploring.

Different early adopters wished assist with concepts, which led to the event of the ideation widget.

Diego additionally recognized that customers wanted extra choices when it got here to producing copy. So, he and the workforce constructed three new AI inputs; Rephrase, Summarize, and Develop.

Diego’s the primary to share how priceless person suggestions has been in creating Buffer’s AI Assistant.

“We actually cherish it, strive our greatest to all the time be very open to concepts and strategies, and act on buyer suggestions.”

Publish-launch success

The AI Assistant launched on February 15th, 2023, and was one of the vital profitable launches in Buffer’s historical past.

Since launching, Buffer’s AI has assisted in creating over 33,300 posts for 8,850 customers. That’s 2.8 million AI-generated phrases created in Buffer.

I requested Diego what side of the launch he’s proudest of. He mentioned he’s most pleased with the way it has turn into an integral a part of the workflow for customers who weren’t assured of their writing talents.

“It’s implausible to see not solely Buffer prospects, however even my Buffer teammates, use AI to be extra energetic on social media. A number of colleagues who weren’t assured writers now use AI as an integral a part of their workflow—they’re constructing an viewers due to it.”

The AI has been used for every thing from brainstorming concepts to refining and sharpening captions for social media posts. Buffer workforce members reminiscent of Maria, Amanda, and Mitra have all benefited from the AI Assistant in varied methods.

Whereas Maria makes use of it to provide you with her private LinkedIn posts, Mitra makes use of it to generate new concepts for the Buffer TikTok account.

What’s subsequent?

I requested Diego what was subsequent for Buffer’s AI Assistant.

Personally, I share a couple of dozen new concepts for Buffer’s AI every week, so I used to be notably eager about Diego’s response.

Diego shared that he’s very excited for the potential of AI to assist creators and small companies save time on duties which can be much less about creativity and extra about formatting, enhancing, and scheduling content material.

Buffer has already saved numerous hours for small companies around the globe, and Diego envisions the AI Assistant will quickly be capable to assist with content material creation as nicely.

This could allow companies to spend extra time specializing in their tales and connecting with their audiences.

However, Diego acknowledges that there is likely to be issues about AI-generated content material missing a human contact or turning into spammy.

“Highly effective instruments can all the time be misused, and I feel new AI developments won’t be the exception provided that that is merely a mirrored image of human nature.”

Concerning upcoming options for Buffer’s AI Assistant, Diego states that the main focus will probably be on enhancing the person expertise to make the expertise extra accessible to non-technical customers.

This consists of refining the interface, offering a extra human language, and optimizing for particular buyer wants.

As all the time, you may control precisely what’s coming subsequent in Buffer’s transparent roadmap.

And eventually … what’s Diego’s favourite immediate?

Thus far, customers have given Buffer’s AI Assistant over 35,000 totally different prompts (I listed my top 38 here).

However I wished to know what Diego’s favourite immediate is. Right here’s what he mentioned …

“Life’s humorous is not it?”

Little bit of a bizarre one, isn’t it!

However Diego mentioned he enjoys seeing how the AI responds to open-ended questions and observing its default “temperature” and humorousness.

Out of curiosity, I plugged that into Buffer’s AI and acquired this response.

It’s not one thing I’d submit on LinkedIn, but it surely’s fascinating, nonetheless!

---

Diego would like to know what you consider Buffer’s AI. In case you have suggestions, please do send us a tweet and let us know.

And if you wish to check out Buffer’s AI Assistant for your self, then get started today, it’s free!)