AI Assistants in Practice: Three Common Workflows

We have identified recurring workflows across industries and companies, for which we are developing generative AI assistants.

According to a McKinsey report[1], tasks that take up to 60–70% of knowledge workers time daily could be automated using current generative AI technologies. This applies to several roles such as marketing, sales, customer service, and product development.

While opinions may vary on the true capabilities of generative AI and the actual time savings be smaller, the speed and efficiency it enables will significantly impact every company’s competitiveness and cost structure. It’s an opportunity no business can afford to ignore.

At Aines, our goal is to help every company realize these benefits as quickly and cost-effectively as possible. That’s why we’ve identified recurring workflows across industries and companies, for which we are developing generative AI powered assistants. This way, companies don’t need to invest in or maintain the necessary technology for these tasks, but can instead focus on their core operations and invest in generative AI development in areas specific to them, such as their own products and services.

Companies don’t need to develop and maintain the necessary technology themselves to automate common recurring workflows. They can focus on their core business instead.

Before founding Aines, we conducted a market research online and among Finnish companies. From our findings we identified recurring workflows across multiple industries and companies that are ripe for generative AI adoption. The three most prominent were:

1. Responding to requests for quotations (RFQs)

2. Purchase order processing

3. Handling exceptions in supply chain operations like P2P and O2C

   
   

Below is a brief overview of how Aines AI assistants can help with each of these.

Responding to Requests for Quotations (RFQs)

The AI assistant identifies the customer’s needs and preferences, whether the request comes via phone, email, Excel, or even WhatsApp. It considers industry- and company-specific parameters, common questions, and the customer’s intent—whether they are looking to order specific products, compare alternative products, or seeking a solution to a specific situation or problem.

The assistant then outlines a suitable proposal based on the identified need, the company’s offering, and other relevant data (e.g., order terms), prices it, and drafts responses to the customer’s general questions. Finally, it writes the quotation according to given instructions or templates, in the desired channel and language. If it encounters difficulties or the quote exceeds a predefined threshold (e.g., in value), it flags it for human review.

But the work doesn’t stop there. In the background, the assistant adds the customer and lead to the CRM, collects data on customer needs and questions that can be later used by functions like marketing and product development, and can even create a reminder in the salesperson’s calendar to follow up on the quote.

The greatest value comes from the assistant handling repetitive information gathering and preparation, speeding up quotation process and allowing sales to focus on what differentiates your company and on building customer relationships.

The competitive advantage comes from AI solutions that accelerate and streamline entire workflows.

---

Purchase Order Processing

The AI assistant developed for purchase order processing first extracts general customer information, delivery instructions, references to previous quotes, instructions for order confirmation etc. It then identifies individual order lines and related data, comparing them to the company’s product information to match them precisely to products and services for entry into the ERP system. Again the assistant can take into account industry and customer specific attributes like product and service features.

It also extracts prices and discounts stated in the order, compares them to price lists and/or contracts, and verifies their accuracy. Finally, it performs quality assurance to ensure all mandatory information defined by the company is provided. If not, it requests human review so that customer service team can fill in the blanks and/or sends a message to the buyer to request missing information.

Once everything is confirmed, the order is processed and sent to company’s systems for production and delivery. In addition to removing repetitive and error prone manual tasks, the assistant speeds up the “order to cash” cycle, decreases risks for human errors, and gives production and supply chain teams more time to respond to demand.

---

Handling Exceptions in Supply Chain Operations

When everything runs smoothly, current ERP and similar systems can handle many tasks automatically. But when there’s a hitch—for example when unit prices in invoice don’t match with agreed prices, or quantities don’t match with the related goods received note (GRN)—humans usually need to intervene.

The AI assistant designed for this purpose can internalize your company’s Standard Operating Procedures (SOPs) and either assist with or automatically handle these exceptions. This may include, depending on the respective SOP, requesting additional information from the supplier, confirmation from the buyer, and ultimately releasing the payment in the company’s systems.

Again this frees up time for supply chain and finance personnel to focus on more value-adding tasks and reduces errors and unnecessary costs caused by haste.

---

Our approach at Aines

All our AI assistants are tailored to match our customer’s business and offering, and integrated as a seamless part of the existing workflows. This way, individual employees don’t need to learn how to use the assistants or develop new prompting skills to get the benefits, and adoption of the assistants is faster than with chat-based solutions.

All data we process and produce, including metadata generated by the assistants, such as questions extracted from RFQs, belongs to the customer and can be delivered to their data lakes and warehouses. And no customer data is used in improving Aines’ services or the underlying large language models.

AI is the new business element. Adopting a general chat assistant is simply not enough—the real competitive edge comes from solutions that accelerate and streamline entire workflows. Aines can help with this.

---

[1] McKinsey, The economic potential of generative AI: The next productivity frontier, June 2023