Will Ai Replace Partner Management?

in on

I don't think anyone in tech would disagree with the idea that Ai is here to stay, and it will get better and better at performing many of the tasks humans have had to spend time on. 

The argument of whether Ai tools, models, and systems will "Replace" entire operations, and event managerial-level jobs, in the relevant near future, is more difficult to have. 

There is a lot of content on how Ai will be used in Sales. But not a lot about Ai in partnerships... So here goes.

In this article, I will try to prove what Ai will soon replace, and what that means for partnership teams.

To start, here are some typical tasks a "Partner Manager" performs today:

  • Onboarding calls
  • Email updates
  • Research and discovery
  • Outreach
  • Putting together partner training videos and materials
  • Prepping & hosting events
  • Running reports from CRM and PRM data
  • Sending partner payouts
  • Trying (and often failing) to work between sales and marketing departments to get partners more involved...

Underneath each of these tasks, there are many trivial tasks that are starting to be prepped by Ai: 

  • Handling the rough draft reports from CRM and PRM data (e.g. HubSpot has a built-in Ai to generate reports)
  • Generating LinkedIn reply and email drafts (sometimes the full emails... although this is not recommended).
  • Generating better-together storylines between two partners for partner acquisition and enablement content.
  • Finding of intent signals and then finding lookalike partners (tools like Owler or Ocean.io).
  • Scoring partner health and leads (using Gong's Ai or a custom rule set between multiple tools).
  • Enablement and nurturing content creation (ChatGPT prompts).
However, there are many aspects of the partner managers role they will continue to be vital in achieving that I for one do not predict Ai replacing any time soon. 
"According to a report by Gartner, by 2025, 80% of sales organizations will have shifted to AI-driven processes, increasing productivity by up to 30%. However, the same report emphasizes that human sales skills, such as storytelling and relationship building, will remain critical for success." Read about that here.

Here's my list of where a human will be necessary to run a partner program:

1 -  Anything involving verbal communication between people - 

  • The sale itself. Partnerships require someone at some stage to convince another human being to either get in their partner program or stay in it. There's a saying in tech "Tools are bought, but Platforms have to be sold." The same goes with partnerships - they are not simple decisions and require a lot of trust between the in the company. 
  • Taking calls with needy partners. While Gen Z and younger are more used to receiving only text and video training, there will always be those who "need" a human on a call to feel they are supported with their needs. 
  • IRL events. The connection between two people can make and retain a partnership between two orgs. When people feel like they "know" someone and have shared experiences with that human, their partnership is far stronger than if they have only ever read their messages or chatted with them virtually. Getting partners to invest the time at these IRL events creates longer-lasting partnerships that Ai could never replicate.
"The people are your partners." - Alex Glenn, CEO of Partnerhub®

2 -  Verification on anything Ai-generated before another person receives it -

  • Human verification on any lead-facing text. If you have never received a cold email clearly written entirely by Ai... It's painful to read. This may be because whatever language model (LLM) it was trained on is not include new cold email strategy or examples... Or that it's simply too new to know why it's bad. But I can guess what's happening in this example especially... The person simply used automation on top of OpenAi which brings in search results for the keywords (like my name, company, etc...) and our company may not rank highly for the general term "Partner Hub". So, the data points it included in this email are simply from a high-ranking website it found: 

    3 - Maintaining compliance, and working around regulations - 

    • Google will soon be able to detect emails written by Ai and notify the user. This regulation will mean teams reliant on Ai-written email copy will have to rethink their strategy. That's something only a human can do.
    • The EU has already put in heavy regulations on tech orgs who leverage Ai. This means those tech companies may not even be able to use Ai to solve some of these operations. 
    • Larger orgs like Google have been warning team members against leveraging Ai models for internal work - even prohibitting it to avoid lawsuits or leaks. Partner teams in these orgs will not have the freedom of leveraging Ai to run operations.

    There are many more, but these 3 alone should be enough to make my point. 

    Ai, however valuable for replacing mundane tasks, is far away from being able to replace anyone on the partnerships team, let alone a manager. 

    Regulations and compliance measures put on Ai will also play a major factor in it's use across the board, but especially for larger tech orgs

    Geo-political regulations - 

    The EU is regulating Ai usage heavily for large orgs especially like they did with cold outreach. This forces salespeople to learn and develop ways to work in the system.Further, it means companies will be staffing full-time roles of experts in sales compliance under these new Ai rules. If they don’t, they’ll risk fines and lawsuits.

    Internal rules - 

    Google, for example "warned its employees about using chatbots, like ChatGPT, Bing, also including its own Bard, at work." This is because whatever they put into those channels with Ai attached, goes into the model regardless of it's proprietary nature. Read more about that case here.

    Regulations on the models themselves - 

    There is starting to be good regs around companies creating models for their industry, but more will come as customers get dupped into a purchase decision based of an "unbiased" Ai which happened to be built and trained by a company selling in that sector. Imagine if Ford created and trained an Ai called "AutoAi" and started marketing it to car shoppers without mentioning it was developed by Ford... And, of course Ford would provide the model far more research and reviews and information on Ford models. Why wouldn't they "forget" to include some small new car manufacturers data in that model? This results in millions of dollars in sales of Fords one year. And then a class action suit against the company when those users of AutoAi find out it was pushing them to Ford by design. 

    How usage of Ai will that change in 2024?

    Today, teams are in an experimental and very superficial stage of leveraging Ai.

    They'll use ChatGPT to generate content, summarize topics, and create rough draft emails and messages to their partners... Ai-enabled tools for generating reports, reviewing partner calls, enriching data...

    Mainly, tasks to save time.

    Fast forward to 2024.

    Now, the overeager individuals letting the Ai go into customer-facing processes are starting to get backlash because they did not run the human checks and balances...

    The organizations then have to have strict oversight on how their departments leverage which LLMs for which purposes so the employees don't generate an Ai-written campaign, or Ai-written code, that gets the company sued (i.e. the Ai code includes protected code that was available in the model). This is already happening with the case of Google and their own Ai Bard.

    Larger enterprises will start to develop their own proprietary trained language models to be deployed in each department based on their customer profiles and industry or geo-specific compliance regulations. They'll have control over what goes in and out.

    Teams next year will still be able to leverage Ai to run their tasks, but much more will be regulated, internally and externally, and at stake if they go against those regulations.


      Just for fun, after I wrote this, I asked ChatGPT the same question and here's what it said: 

      Artificial intelligence (AI) has the potential to significantly impact the field of sales, but it's unlikely to replace sales management entirely. Instead, AI is more likely to augment and enhance various aspects of sales management. Here are some ways in which AI could influence sales:

      1. Data Analysis and Insights: AI can analyze vast amounts of data to provide valuable insights into customer behavior, preferences, and trends. This information can help sales managers make more informed decisions about strategy, pricing, and targeting specific customer segments.
      2. Lead Scoring and Qualification: AI-powered algorithms can analyze leads based on various criteria, helping sales teams prioritize leads that are more likely to convert. This can save time and improve the efficiency of the sales process.
      3. Predictive Analytics: AI can be used for predictive modeling, forecasting future sales trends, and identifying potential upselling or cross-selling opportunities. This helps sales managers make proactive decisions and allocate resources more effectively.
      4. Automated Communications: AI-driven chatbots and virtual assistants can handle routine customer inquiries and tasks, freeing up sales teams to focus on more complex and relationship-oriented aspects of their roles.
      5. Personalization: AI can analyze customer data to provide personalized recommendations, content, and offers. This level of personalization can enhance the customer experience and increase the likelihood of a successful sale.
      6. Sales Coaching and Training: AI-powered tools can analyze sales calls and provide feedback to sales representatives, helping them improve their communication skills, identify areas for improvement, and enhance their overall performance.
      7. Workflow Automation: AI can automate repetitive tasks and streamline workflows, allowing sales managers to focus on strategic decision-making and relationship building rather than administrative tasks.

      While AI has the potential to automate certain aspects of sales management, human qualities such as empathy, relationship-building, and strategic thinking are likely to remain crucial in sales roles. AI is more of a tool to assist and enhance the capabilities of sales teams rather than a complete replacement for the human element in sales management. Successful integration of AI in sales requires a thoughtful approach that combines technological capabilities with human skills and intuition.