The Partner Case for Dogfooding AI
Everyone has been talking about AI since ChatGPT was released, but even for years before that, AI has caused plenty of buzz in our industry. But where exactly is the money in AI?
On the cost side, it's quite obvious. Many billions of dollars are being invested in AI by the cloud hyperscalers, different platform companies and everyone else. On the income side, however, it's a little less obvious.
Last year, the four big tech giants increased their spending in AI by a staggering 63 percent. Microsoft, Alphabet, Amazon and Meta reported combined capital expenditures of $246 billion in 2024, up from $151 billion in 2023. Their spending is projected to exceed $320 billion in 2025. That's a lot of money, creating lots of expectations.
At first, the stock market reacted positively to these investments. Now, though, it is becoming impatient: Earnings are not growing at the pace it had hoped. The hyperscalers all invented new and wonderful AI services that customers all over the globe love and use, but the bump to their profitability is smaller than investors wanted -- and the stock market is punishing them for it.
There's a very important exception, however: Meta, the company behind Facebook and Instagram. Meta engineered a successful path for embedding AI in its services, and it found the formula to improve ad targeting on Facebook and Instagram, which is great for its profitability. Meta doesn't sell AI tools; instead, it sells its two flagship services, Facebook and Instagram, and has used AI to improve them.
This validates my thinking that the low-hanging fruit for AI profitability is improving internal processes and existing products, and reducing your cost base. Using AI to gain better internal efficiencies is something you can control and start immediately. Not since the advent of PCs have you been able to boost the efficiency of any company as significantly.
There are exciting new services based on AI and a growing number of vendors investing in them. But what will deliver a positive bottom line return today -- never mind tomorrow -- is using AI internally. Even if you're not as big as Meta -- and you're probably not -- there are a number of things you can do internally with AI to drive increased profitability.
Embed AI in your service desk so you need fewer people to handle a larger volume of support tickets -- and use AI to reduce the number of tickets. This means you can grow faster and easier than before because you don't need to hire in the same pace. You might even be able to redistribute your talented people to other areas of your company.
Use GitHub Copilot for software development. AI will not be able to create every line of code needed, but it will probably give you 80 percent or so, and that is a game-changer in terms of cost and time to market. Replace all, or some, of your junior developers with GitHub Copilot and keep your senior experts.
Marketing with AI is revolutionary. Experiment until you find the right tool for creating all the assets that you need. You'll be able to do so much more in-house and outsourcing will be rare. Your younger marketing people might be the best ones to empower and to encourage to use AI. Start experimenting!
Classify and organize your data so your internal AI services read the data that is up to date and confidentiality is respected. There are a few great tools out there for how to do this.
Personal Copilots will make each employee more efficient. There are more than 100 available from Microsoft today, and even more from other vendors. Make sure that everyone has a license for Microsoft Copilot and let people freely experiment. Create policies and build awareness for what data should not be shared externally so that your confidential data is not being used to train publicly available large language models (LLMs).
Add functionality driven by AI in the products that you sell. If you're an ISV, adding AI driven features might make sense. If you're more of a services company, explore if AI can make your services better for your customers and less expensive for you to produce.
Build agents that takes care of various processes, and let the agents interact with each other as needed. Agents will be used for various roles, and they can interact with people or with other agents. Don't be surprised if a new title in your company's organizational chart is Agents Director (you heard it from me first).
Aim to reduce friction inside your company with the use of AI. Increase speed in everything that you do and always ask what else can be made better with the use of AI. Make sure this spirit permeates all levels in your company.
And here's the added bonus: When you use AI internally for your own company's benefit, you'll not only enjoy higher profitability, but you'll also be in a terrific position to help your customers achieve similar positive outcomes. You'll be tremendously trustworthy when you tell your customers about how you use AI internally, and the conversation with customers will be a walk in the park.
Good luck and please share your success stories with me!
Posted by Per Werngren on March 25, 2025