How AI is transforming RFP responses (and the GPT future)

26 May 2023 | 15 Shares

Source: Avnio

Since its debut in November 2022, ChatGPT, a large language model (LLM) powered by OpenAI’s GPT algorithm, has garnered significant interest. In its latest commercially-available form, GPT-4 builds on the foundation of its predecessor, GPT-3, having utilized larger datasets and increased computational power during its training cycles. Models were fed vast amounts of data from various sources, including scientific archives, books, reports, forums, and news coverage. That’s enabled OpenAI to develop increasingly accurate statistical models of language and knowledge.

The GPT series of LLMs are highly effective due to using attention mechanisms to predict the next word in a sequence. GPT-4 has a context length of 32,768 tokens (equivalent to over 50 pages of text), enabling it to summarise lengthy documents and engage in thought-provoking discussions.

API access to these advanced language services now has the potential to transform enterprise software, including – the subject of our focus today – RFP automation tools. That’s transformative because AI already offers the ability to address complex RFP documents and helping guide companies responding via RFP to tenders.

The extension of existing capabilities with LLMs like ChatGPT4 is likely to disrupt how companies respond to all types of tender documents. That will happen not just through the emergence of more powerful models but in how more extensive data is collated locally, and the complete dataset is quizzed using natural language.

Source: Avnio

Source: Avnio

Knowledge Library

Pre-trained models can be fine-tuned with corporate data – keeping the information sandboxed for security – allowing LLMs to respond to not just general questions but highly detailed business queries that pertain to a single company. Even answers to questions not experienced before can be created by AI routines. This capability, which is still uncommon, already delivers huge productivity gains. For example, Avnio – a provider of RFP automation software – finds that using AI to generate executive summaries, cover letters, and autocompleted business details can reduce the time its clients spend on RFP, security questionnaires, and other company requests. The majority (~90%) of questions in tender documents can be answered automatically using a combination of AI algorithms that mine pre-taught models combined with data local to the responding company.

Its users can take an RFP, including attachments in any file format, and drop the electronic documents into a Salesforce-native workflow. The application uses language recognition to understand digitized paperwork and structure the various information fields into categories. The tool can identify whether entries relate to security or compliance topics, legal considerations, and queries on company structure, to give just a few examples.

Linking the incoming requests to an organization’s knowledge library allows the RFP automation platform to generate a first draft in seconds. A company’s library of RFP question responses will build rapidly, thanks to AI-powered contextual interpretation and response creation.

Users already have the option to edit and finesse RFP responses generated by AI right in Salesforce, plus they can raise queries through an assistant and make quick edits via Slack and Teams. Smart automation can only improve win rates as the AI platform in the back end improves over time.

The qualification rounds

Automation tools give bid writers more time to focus on the proposal’s key writing and editing stages and put their organizations ahead of the competition. The autocomplete function in the Avnio tool is more than just a time-saver. The answers provided by humans as responses to the RFP document are also analyzed by the Avnio platform to determine how accurately the answers match the questions on the RFP (or other input documents) – shining a light on the company’s propensity to win new business.

Wider data resources from Salesforce inform the RFP response process with information like specific consultants’ win rates and contract value, region, contract type, and competitor data: in fact, any way that data can be sliced and diced.

The system’s use of machine learning can help determine whether the tender is a good fit when the company’s capabilities and resources are considered. This is output as a ‘qualification score’ in a percentile, allowing Avnio users to yes-no RFPs much more quickly, and letting them concentrate on projects where win rates are predicted to be much higher.

For companies considering high numbers of RFPs and other information requests, this scoring step allows them to quickly prioritize which questionnaires to place at the top of their to-do lists. And by optimizing their RFP processes, users will be more likely to win more business.

Companies can update their Salesforce-based knowledge libraries with the information and insight gained from submitted proposals. And while firms will, naturally, be keen to win rather than lose bids – the feedback is useful regardless of the outcome.

Native to the Salesforce environment, Avnio’s platform, which is available on the AppExchange, can pull in information from Salesforce objects and use those databases to identify patterns and supply supplementary data that enriches responses to RFPs.

Source: Avnio

Source: Avnio

Wider organizational benefits and the Future of Automation

As GPT comes online, data analysis will become available via natural language interfaces. Queries around underlying or hidden patterns to RFP success rates could be couched in simple queries: “How is our record of success in the Sustainability section of RFPs in this sector we’ve submitted in the last three years?”

The power of GPT4 means that any company leveraging it in commercial contexts can draw from the wide learning corpus that has trained the model, including academic papers and statistical analysis. Therefore, as the technology comes online in the RFP space, queries could compare, for example, the company’s resources and RFP responses against those of competitors. Insights from the correlations of massive data sets will be just a natural language query away.

Present-day Next Steps

Bid managers, bid writers, tender management teams, and heads of deal structuring will all benefit from the focus on what’s succeeded in responses to date and, more to the point, why. Insight into where the company is missing out will improve win metrics rise, thanks to unearthed trends that the Avnio AI-powered RFP automation platform delivers.

Avnio and Salesforce recently discussed RFP automation in a live webinar with a key focus on the possibilities of GPT4 and the company’s roadmap toward its integration into RFP processes. You can catch it here.