Performance management has been and always should be person centric activity, your employees, day by day, month by month are performing for you. With you leading regular employee conversation both ad hoc and formal. However, there are places for AI to augment the human lead and approach. Yes, AI can be used prepare agendas, identify follow ups, record and transpose the actions and outcomes of performance conversations. We know that is bread and butter for artificial intelligence. But how can leaders use AI more meaningfully within performance management world? How can leader start to recognise there is a balance to be had, in managing performance, using available tech and maintaining the critical human touch?Â
This article serves to stimulate individual and organisational thought around the topic.
There are several key areas where AI can be a useful tool in supporting leaders in their performance management activities:
Scheduling Performance Management
Performance management cycles and the actual appraisal process are predictable, which feeds into AI’s natural domain. Organisations can implement AI platforms around performance management that maximise these cycles and processes leading to increased efficiencies across the board by scheduling conversations, prompting for peer feedback, then collating and analysing this. Straight away this brings performance management front and centre, without the need for HR to send out continual reminders. By scheduling conversations it also prioritises and brings focus to the employee being front and centre.
Objective Setting
In my experience many leaders struggle here. They struggle to tease out work for team members from the bigger picture. They then struggle to translate this work into meaningful SMART objectives for measurable outcomes. AI can help to interpret, process and organise key business information into workable direction for employees. Or at least give and indication, a starter for 10. Where there is a sizeable piece of work to translate to objectives, artificial intelligence can identify natural and aligned smaller chunks of components of deliverables and potential objectives. Enabling leaders to more effectively translate into SMART objectives.
AI and SMART objectives – yes in part, but the objective will only be as good at the prompt which is created in the first place, which actually might take more time than creating the objective itself! Perhaps better to learn how become competent at using SMART as a leader? Also, only the leader’s unique employee insight will be able to accurately determine if the objective is achievable for the employee concerned given their experiences and current capacity. Also, if it is relevant for their current role and business need.
So, any SMART objective produced by AI might end up being generic in nature, and need some fettling.The ability to update objectives after role or work shifts is key to maintaining performance, however the frequency that this happens by employees/leaders can be mixed. However, if a system can recognise these shifts it can either prompt for new objectives or help identify new ones. This can help maintain effectiveness.
Performance behaviours
Objectives are made up of the WHAT (the actual objective) and the HOW (the enabling behaviours around the objective). Competencies and behaviours can be identified, mapped explained readily through AI transposing organisational competencies onto objectives. Thus, enabling the conversation around objectives to be more meaningful.
Patterns and trends in performance
Where employee work and deliverables produce, utilise and process data artificial intelligence can be used to track performance through spreadsheets. That is a natural home of artificial intelligence – analysing data. Trends, patterns and themes provide important insight for line managers. This insight combined with their experience of employees and employee conversations provide a more holistic perspective on performance. This has the potential of leading to providing predictive performance data, which has its pros and cons.
It is already clear that performance management cannot and should not be simply ‘handed over’ to AI, the human leader contributions and enablement are critical. These conversations also will further enable mutual understanding of both the objectives and the wider influencing contexts, as well as the expectations that surround them.
Unilever use an AI platform that focuses on feedback loops. Employees now receive ‘real-time insights into their performance and development’. This has been positively received by employees. This also reduces ‘recency bias’ where leaders often can just focus just on what has been delivered ‘recently’ around performance, as well as reminding employees what they achieved several months ago!
Coaching questioning
Using AI to help formulate open based questions around performance management, or to test the validity of their own questions, can be a useful tool. But it shouldn’t be used to replace leader-based questions creation, but to perhaps consider different angles to questions, or to stimulate more thoughts and ideas for the leaders. Coaching questioning is a powerful everyday communication tool that leaders will benefit from being skilled in. Simply handing this over to AI will lead to leadership deskilling.
JPMorgan Chase now reputedly use chatbots to help employees draft their performance reviews.
Employee development
By developing SMART(er) objectives through AI, at the same time it can potentially provide insight into potential development areas around these objectives which might be relevant. Leaders can then test these against their own reports current skills, knowledge and behaviours, to come up with a more complete picture of development needs. Therefore, AI can potentially look at performance and support the proverbial ‘gap analysis’ around employee development. Over time there is the potentiality of personalising development as well as considering wider patterns in team development.
Team performance
Where a team’s objectives are known, AI can help identify and show where individual collaboration can be maximised, or where interdependencies lie. Where access to them data can be sought from AI, as well as patterns and trends, provides a great tool for teams to sit down and discuss team performance. The outputs from these discussions can transform not only team performance but also team dynamics and relationships.
IBM use their IBM Watson to analyse multi-source performance data, detect patterns and highlight potential manager bias during appraisals
Where AI should be only lightly used in performance management
Employee feedback
There is no doubt that AI can produce reams of feedback. Again, though as feedback is highly personal it should be primarily be a person to person, face to face, person activity. Feedback isn’t a one-way activity, it is primarily an engagement activity. Where leaders should have the ability and confidence in in their feedback approaches, language and performance awareness. Whilst AI can be used to identify potential feedback approaches, it should not be used to craft the feedback language.
Any report will readily identify when AI prepared narrative is used, as they will have an idea of their line managers natural language and nuances. Similarly praise and recognition should be led and initiated personally by the leader rather than it being AI led. In reality providing these two cost the line manager nothing in terms of time and effort.
However, to counter this a little. In a Gartner survey they found that nearly 90% of employees believed that algorithms could give fairer feedback then their line managers! So, there is a balance to be struck and the roles of each explored.
Performance management is an informal daily activity and part of the leader’s required personal skill set. Not just predictable monthly 121s or an annual process. The more leaders are ‘hands on’, the more they are gathering information first hand and can make more informed and evidenced based decisions around employee performance.
Poor performance conversations
Where poor performance is identified and conversations needed, yes AI can be used to help prepare for conversations and manage potential responses, however there is huge risk if AI is used to direct how to do the conversation. As this should be primarily a leader-based conversation, utilising their internal skillset. Denegrating people poor performance process and experience to AI, give employees a clear message in terms of the value they are held by their leader and organisation. Employees deserve proper human and leader contact in these matters.
Where performance information is inputted into a platform and AI used to make analysed observations, it should not be used to infer or judge that someone is underperforming, disengaged or not committed. This is because context, personal situations etc will not have been considered in the decision process.
The challenge with AI and performance management
Yes, it is clear AI can be used in multiple areas to support performance management, it has the potential to be really enabling for both employee and leader. However, what AI can’t do is consider the multitude of variables and nuances that also influence performance and people in the work place. Motivation, team issues, individual issues, changing priorities, leadership bias, politics, support, development needs, personal issues etc. etc all can come into play.
It can’t replace those ad-hoc day to day conversations, observations and instincts that inform a leader’s holistic view on performance. AI can focus on producing facts and figures but lacks the insight at present to effectively interpret this with the variables that surround it as well.
Some AI performance AI platforms are now able to analyse communication patterns to determine engagement levels and morale. Whilst this can be positioned as a wellbeing tool – to prevent stress and burnout, how else might this information be used?
There are some leaders who would quite happily ‘handover’ performance management to an AI platform or AI bot, if they could. It’s just not their thing! But in doing so they miss the fundamental roles and purposes of being a people leader. Also, performance management in its entirety presents so many positive opportunities for discussion, planning, learning, relationships, growth and cohesion, that AI can’t.
Whilst AI performance management companies promote the benefits and efficiencies of their platforms, they are not talking about and leaving the ethical implications on the doorstep of the purchasing organisations. Privacy, surveillance, deep data collection, constant tracking, prejudiced outcomes and lack of awareness by organisations of how employee data is being processed, are but a few ethical issues to be considered before implementation.
There is no doubting AI’s place in modern organisations and people leadership. It’s progress within business is inexorable. However, despite the key value add it can bring, it should not be a carte blanche take up. For both individual leaders and organisations, key planning, testing and consideration must be undertaken to understand how to use and apply, not only as part of due diligence but also to increase embracing of the technology and to reduce suspicion.
Nick Howell is a Coach, Facilitator and Author of ‘Great Coaching Questions‘. Working with leaders, teams and individuals to bring change and growth. Using coaching, workshops, 121 development opportunities and programmes to bring personal transformation in the work place.