Stéphan Donzé, Founder and CEO at AODocs – Interview Series

Stéphan Donzé is the founder and executive of AODOCS, a platform for managing original cloud documents that transforms the institution’s content into practical intelligence. Unlike the LEGCY systems that are limited to the basic storage, AODOCS combines strong control of documents and automated workflow, allowing companies to simplify complex operations across industries.

From sales proposals and demand management to engineering and manufacturing, AODOCS AI is used to extract organized data from non -structured documents – widely.

What inspired you to start AODOCS in 2012? Was there a specific problem I saw in managing documents that were not addressed at that time?

This came from the concept of providing business technologies to the consumer (B2C), the pioneer by Google, to the institution. This included cloud -based cloud techniques, without a servant, and automatically expanding technologies that were absent in the Foundation’s program at that time.

At that time, Google was defending applications to adapt to the cloud by moving away from the thick customers and embracing the browser structures. Unlike Microsoft and Amazon at the time, Google’s position was that cloud applications should adopt a completely different structure of hypothesis programs. I saw an opportunity to create something new by combining old and new methods.

The first adoption of cloud technologies in institutions, who adopted Gmail and Google Drive, wanted to extend these benefits beyond e -mail and cooperate with important documents for business. There was a lack of solutions in the market to meet this need, as traditional document management systems were local with old and brown programs. This gap presented an opportunity to build something new from scratch on the structure of the right cloud, and to work with these first adopters.

Many Saas Institutions companies are struggling with the adoption of customers. How did major companies such as Google or Veolia convinced confidence in AODOCS early?

Allowing customers to keep the documents in their cloud storage was an important factor in the convincing institutions. AODOCS run documents, including sharing permissions and business operations, but documents remain within the customer environment (initially Google Drive, Azure, Google Cloud Storage and Amazon S3). This basic feature was reduced by the risk of customers, and made them more comfortable working with a small and young company.

Another major reason was close customer relationships. We have participated in our construction and developed with customers, especially in the early stages. This proximity and interaction in repairing problems, making improvements made of confidence and increasing risk perception.

This was in the end what companies were convinced like Google, Veolia, Ascency, Pinnacol Assrance and others to adopt.

If we look back to more than 13 years of AODOCS leadership, what are some of the most important decisions that led to the success of the company?

The decision to maintain customer documents in their cloud storage is a great factor in our success and remains a major discrimination factor.

The development of the product by staying very close to customers was very important. All that we build is inspired by or verifying health through customer needs. We avoid building features without a clear use of the customer.

Our ability to sort through customer requests to determine the common and repetitive needs that benefit the entire customer base, unlike the very specific individual requests that are best addressed by professional services, is a good thing in.

AODOCS was born in the cloud and is now developed into a behavior -working solution. What are some of the biggest challenges in making this transition?

It is not a lot of transition, but continuity. The generation in the cloud makes Aodocs easy to merge with new cloud -based techniques, including artificial intelligence.

Artificial intelligence models such as ChatGPT and Google Gemini depend on the cloud, making it natural for AODOCS to interact with them and add Amnesty International’s capabilities to our document processing platform.

AODOCS is considered as the “spine” or “assembly line” for document operations, as artificial intelligence is simply “party” or a new “unit” that can be connected to strengthen the statute. We did not expect the appearance of artificial intelligence specified in 2022, but our original cloud structure was designed for expansion and integration, making artificial intelligence adoption a natural step.

How to balance AODOCS automation with human control, ensuring compliance and accuracy without removing human health verification?

We offer customers to use artificial intelligence in different steps in their documents operations. The autonomy decision to present artificial intelligence is eventually taken by customers on the basis of each case separately.

Customers can create their operations for full automation, full human review, or a combination of both of them in any step of artificial intelligence. We recommend identifying easy cases of fully artificial intelligence automation (for example, discovery of lost documents in a request), while artificial intelligence can be used in complex cases to accelerate human review, usually by creating complex documents summaries.

Basically, we give customers 100 % control over the level of human control in their operations by providing training tools.

AODOCS converts its platform to focus on artificial intelligence agents to process documents. What prompted this decision, and how artificial intelligence agents improve on traditional document management systems?

Artificial intelligence agents are a new capacity that can be combined into our current “assembly” documents.

The primary driver to excite us about artificial intelligence agents is their ability to achieve great speed to specific tasks. For simple tasks such as rejecting incomplete documents or accepting low value, directly direct claims, Amnesty International can enable full automation.

For complex tasks, artificial intelligence factors can greatly accelerate the processing by summarizing long documents, allowing humans to review the main information much faster. Artificial intelligence excels in eating large quantities of information, providing brief summaries, and providing a great time for auditors. This increased and efficient speed enhances the value of the AODOCS platform.

Research in the institution remains a major challenge, as companies are struggling to recover relevant documents. How do you address AODOCS ‘AI Solution for discovery better than research tools in current institutions?

AI Chatbots is a significant improvement in the traditional research of institutions by allowing users to ask questions in normal and receive direct answers. However, they have a major defect: they cannot distinguish between documents and good, updated, incomplete or outdated documents.

We face this challenge by ensuring only the use of artificial intelligence agents. The traditional research of users provides a list of potential matches, allowing them to determine the most relevant document (such as the distinction between signed and draft contract).

On the other hand, AI Chatbots provide confident answers but hide the basic search process. This creates an increasing danger to the use of old or incorrect information, as users do not have any vision on documents that Chatbot uses to formulate its response.

Our role is to organize and ensure that all documents at the base of knowledge have been validated correctly and updated. This makes them safe for artificial intelligence agents to use it as information sources. By ensuring the research that works on artificial intelligence depends only on reliable information, we significantly reduce the risk of inaccurate responses.

Even though the obstetric intelligence becomes a major part of document processing, what are the specific techniques or models of ADOCS to improve accuracy and reduce manual work?

LLM models outperform the conversion of the current content, such as summarizing complex documents, translating text and classifying files. These tasks have a very low risk of hallucinations, which is why we use artificial intelligence strategically focus on these activities.

AODOCS takes care mainly from AI to reformulate, reuse and transform the existing text instead of creating a new content. When working with correct materials, creating summaries, translating documents, and content classification, they all maintain a low probability of error. Likewise, our data extraction capabilities – determine the dates of entitlement or major information from documents – are based on the current content to reduce inaccuracy or “hallucinations”.

By accelerating the swallowing of information and supporting human auditors, artificial intelligence capabilities often improve efficiency through an important factor. This creates concrete value while avoiding the risks associated with the content created from artificial intelligence carefully from scratch.

There is an additional feature. AODOCS is smoothly integrated with any LLM, giving our customers the freedom to choose their favorite provider. This flexible structure also guarantees that we can integrate future models quickly when they are available, putting our customers to take advantage of the innovations of artificial intelligence that are inevitable in the coming months and years.

One of the biggest challenges of artificial intelligence in the treatment of documents is hallucinations or inaccurate responses. How does AODOCS ensure that the ideas created of artificial intelligence are trustworthy and reliable?

The key to trust in trust and reliability is grounding. We ensure the use of artificial intelligence to reformulate or convert information from reliable sources.

In the treatment of documents, artificial intelligence is required to summarize or translate current documents, which reduces hallucinations. In AI Chatbots or agents, the system is provided with extracts of controlled documents and is directed to create only answers from those excerpts – LLM must always quote the sources from which their answer was born – with the use of general general knowledge or external information. This prevents artificial intelligence from inventing or providing information that is not explicitly present in documents that have been validated.

With AI Transforming Enterprise Software converts, where do you see AODOCS in the next five years? What role will artificial intelligence agents play in the future of document management?

Our basic work will not change as a document management platform. Our mission remains to ensure the merit of trust and currency, adhere to the proper business process and track managed information.

Artificial intelligence both is a new need for controlled information (to feed the trusted data to AI Chatbots) and the speed of commercial operations that include documents. Amnesty International can accelerate tasks such as summarizing, translating and classifying documents, and making human users more productive.

In the next five years, we aim to provide tools for companies to easily apply artificial intelligence as it makes sense to enhance their businesses. Artificial intelligence agents will play an important role by amplifying and accelerating our basic tasks: providing trustworthy information, ensuring the use of correct documents versions, helping to avoid human errors, and facilitating access to commercial processes. We want to help companies use artificial intelligence effectively and strategy to enhance productivity.

Thank you for the wonderful interview, readers who want to know more AODOCS visit.