Pactum AI

Did you know you can launch autonomous negotiations through SAP Ariba? By sending out our AI agents to make even small improvements to your under-managed contracts, you can see noticeable bottom-line gains. So if your company is a customer of SAP then Pactum’s integration enables you to improve requisition pricing and payment terms with suppliers autonomously, and without leaving the Ariba platform. It’s ideal for negotiating one-off purchase requisitions below $150,000. Learn more here: https://lnkd.in/dEFq_Z3P. Or reach out to Pactum’s Michael Heeks to discuss this further. And watch this space. As autonomous negotiations become a more normal part of every day business globally, there are more Pactum integrations on the way to ensure you can launch them from wherever is most convenient for your business.

Just arrived home from a busy week in New York discussing how companies use AI for autonomous negotiations. I was asked here what is my personal vision or goal for Pactum in five years. Pactum is rather unique and it can be answered in countless ways. What’s more, many people have different flavors of the vision. The vibrant NYC inspired me to go deeper inside one path that I personally also believe in. And the answer ended up being a bit unique. That it’s not mainly about better agreements, higher revenue, reducing costs or anything like that. Or even about a more efficient world or happier business partners. It was about speed and agility i.e. from decisions to material impact without a delay. As our world becomes more digitalised, companies have been extending their capabilities at the front of that process. Usually, it is meant in one direction. Most executives can get very precise real time data about their business operations around the globe. How many items were sold in Barcelona last month? What was the most profitable country for us in Asia last week? But executing even a simple decision such as let’s expand in Europe or focus on quality instead of price, can take many-many years to roll out in a global enterprise. And there are always market changes, new trends, and black swan events for which speed and agility is rewarded. So, one of my visions for Pactum in five years is that there is a company who has experienced the power of fast business execution via autonomous negotiations such that a high level decision will have a desired substantial impact on the overall business. And of course, I also assume that they would love it and would want more of it :) For me, the world of B2B commerce is sometimes difficult to personally relate to. But it’s the same trend I see evolving in our home life too as AI agents increasingly act on behalf to make purchases. If I decided to buy a new car for my family, I would want the decision to be quickly executed by a software system that already knows a lot about me, what I value, how much is worth spending, what features I desire, etc. I would then want it to give me a few offers for different cars that it was able to automatically choose and describe to me, having already negotiated the final offers. We are still far away as we are not large enough with any of our customers yet but 5 years is an absurdly long time.

Companies shouldn’t rush to use AI they don’t understand, advised Kristjan Korjus on Nasdaq TradeTalks this week. This is especially the case with large language models, he elaborated. Kristjan, who has a PhD in machine learning, is our Chief Technology Officer, and has written numerous academic papers on AI. Before co-founding Pactum, he worked as Head of AI at Starship Technologies, the leader in autonomous deliveries. Kristjan explains that AI offers incredible value, but has drawbacks too. Understanding its limitations is key to using it effectively and responsibly. Here’s some of Kristjan’s bullet points for where large language models should be used: - Where users input unstructured text - Where users allow it to make mistakes - Where users understand it can act inappropriately - Where users can have access to its "inner-mind" For these reasons, large language models are most suitable when integrated into an internal tool. Or, if external, in cases where those external users understand that additional work may be required and when there is no proprietary data fed into the model. As a good example, there are tools based on large language models that help developers write code. But the developers know they must still check that code. As a bad example, some companies use large language models for customer service bots. But it’s not a customer’s job to verify what customer service says to them. This mismatch in expectations has resulted in all kinds of fails, such as when those bots have hallucinated special offers. So, how do we use large language models at Pactum? For autonomous negotiations, we use large language models in negotiation preparation. They offer a huge amount of value in tasks such as finding suppliers, sifting through market data, and modelling negotiation flows that can be constantly A/B tested. We don’t put them in control of live discussions with suppliers though. We use our explainable, rules-based AI for that. It means our AI agents will act entirely within the parameters set by our clients and it means their suppliers know they are negotiating with AI agents that can offer them good deals, keep their word, and have the authority to close those deals. That’s what makes it autonomous for our clients.

🎉 Pactum has been selected by Tyson Ventures as “one of the most promising innovators working towards a better food system”. The venture capital arm of Tyson Foods invited innovators and entrepreneurs from around the world to suggest breakthrough solutions that can help Tyson Foods foster more sustainable practices, enhance transparency, create new efficiencies and ensure the highest standards of quality across the unique protein supply chain. At a final pitch stage at Tyson HQ, Pactum’s Nick Tomassetti and Jake Filarski (who you can spot in the picture below) explained how using AI for autonomous negotiations can create a stronger and more agile supply chain in which all sides benefit - including suppliers and customers. The panel was so impressed that Pactum has now been selected among five startups “to engage in continued conversations within the business, paving the way for further high-impact opportunities”. Every day, more enterprises are discovering the benefits of autonomous negotiations and we’re proud to partner with forward-thinking companies to bring those benefits to the world.

It's been some years since Pactum participated in startup competitions. We did an exception here with Tyson. Pactum was selected one of the 5 winners out of 100+ startups. Good to know we still have it. And great to see how Pactonians are all like founders - ready to pitch the startups on the stage. Well done, team!

Bienvenue dans l'équipe, Alexandre Glais! As we bring the benefits of autonomous negotiations to the world, we need to talk about it in ever more languages. So, in case you’re wondering, it’s ‘négociations autonomes’ in French. Alexandre has joined us as a French-speaking Operations Support Specialist - although we’ve already heard him switching between multiple languages to support our diverse community of practitioners in autonomous negotiations. Pactonians are a talented bunch. Alexandre has already made a great impact and we’re excited to have him onboard as we scale globally.

As a CTO and Co-Founder of a chat-based autonomous negotiation company and also as a Deep Learning veteran, I've been constantly thinking about where to use and where not to use the Large Language Models. I'll be in New York on Monday - come join me at the seminar and/or meet me before or after. On Tuesday, I'll be going to Nasdaq TradeTalks to continue the discussion.

Salary negotiations can be tricky. But what if your employer sent an AI agent to negotiate your contract with you? Here’s Grace Spellmann, MS, PHR talking about how it can be done effectively and responsibly in an interview just broadcast by Voice of America. We know that job applicants tend to be weary of any AI in the hiring process - and rightly so. There’s a lot of bad examples out there of how not to do. AI shouldn’t just be a time saver for the employer. It should add new value for both the employer and the applicant. Well, here at Pactum, we’ve got the technology to find better deals for both sides through autonomous negotiations. So we’ve already experimented with the use of that technology internally for job contracts, ensuring that both sides get more of what they want and better long term outcomes. Some of us (including the person writing these posts) are working on contracts that were concluded with our own AI. It’s not a service that Pactum offers commercially - because we are focused on helping companies conduct autonomous negotiations with their suppliers - but it does work just as well internally to augment the process in HR too. As Grace explains, “the bot doesn’t care how many times you go back and forth with it. It wants to find you a good deal that makes everybody happy.”

Today at Tyson Foods, Pactum’s Jake Filarski and Nick Tomassetti did a great pitch for how autonomous negotiations can improve the way our food gets from farm to fork. 👇

Pactum CTO Kristjan Korjus will be on Nasdaq TradeTalks next week - but anyone in New York can first meet him at our procurement meetup. 👇